# API Tokens
Source: https://docs.meetzy.es/account/api-tokens

Create and manage secure API tokens for programmatic access to Meetzy's platform

API tokens provide secure, programmatic access to Meetzy's platform without requiring user credentials. These tokens act as secret keys that authenticate your applications and services.

<Warning>
  Never share API tokens or expose them in client-side code. Treat them like passwords and store them securely in your server environment.
</Warning>

## Accessing Token Manager

Navigate to your account settings to manage API tokens:

<Steps>
  <Step title="Open Account Menu">
    Click on your profile avatar in the top-right corner of the dashboard
  </Step>

  <Step title="Select Account Settings">
    Choose "Account Settings" from the dropdown menu
  </Step>

  <Step title="Access Token Manager">
    Click on the "API Tokens" tab to view your token management interface
  </Step>
</Steps>

## Creating a New Token

<Steps>
  <Step title="Click New Token">
    Click the "New Token" button in the token manager interface
  </Step>

  <Step title="Enter Token Details">
    Provide a descriptive name for your token (e.g., "Production API", "Analytics Service")
  </Step>

  <Step title="Set Expiration Period">
    Choose an expiration timeframe:

    * **30 days** - Short-term testing
    * **90 days** - Medium-term projects
    * **6 months** - Long-term integrations
    * **1 year** - Annual renewals
    * **No expiration (4 years)** - Permanent access
  </Step>

  <Step title="Generate Token">
    Click "Create Token" to generate your new API token
  </Step>

  <Step title="Copy Token Securely">
    **Important**: Copy the token immediately. For security reasons, the full token value will never be displayed again
  </Step>
</Steps>

<Info>
  Choose expiration periods based on your security requirements. Shorter expiration periods are more secure but require more frequent renewal.
</Info>

## Token Security Best Practices

### Storage Guidelines

* Store tokens in environment variables, not in code
* Use secure secret management services in production
* Never commit tokens to version control systems
* Rotate tokens regularly based on your security policy

### Access Control

* Create separate tokens for different applications or environments
* Use descriptive names to easily identify token purposes
* Revoke unused or compromised tokens immediately
* Monitor token usage through API logs

### Environment Separation

```bash theme={null}
# Development environment
MEETZY_API_TOKEN=your_dev_token_here

# Production environment  
MEETZY_API_TOKEN=your_prod_token_here
```

## Using API Tokens

### HTTP Headers

Include your API token in the Authorization header of your requests:

```bash theme={null}
curl -H "Authorization: Bearer your_api_token_here" \
     https://api.meetzy.ai/v1/assistants
```

### Code Examples

<Tabs>
  <Tab title="JavaScript">
    ```javascript theme={null}
    const headers = {
      'Authorization': `Bearer ${process.env.MEETZY_API_TOKEN}`,
      'Content-Type': 'application/json'
    };

    fetch('https://api.meetzy.ai/v1/assistants', { headers })
      .then(response => response.json())
      .then(data => console.log(data));
    ```
  </Tab>

  <Tab title="Python">
    ```python theme={null}
    import os
    import requests

    headers = {
        'Authorization': f'Bearer {os.getenv("MEETZY_API_TOKEN")}',
        'Content-Type': 'application/json'
    }

    response = requests.get('https://api.meetzy.ai/v1/assistants', headers=headers)
    data = response.json()
    ```
  </Tab>

  <Tab title="cURL">
    ```bash theme={null}
    curl -X GET \
      -H "Authorization: Bearer $MEETZY_API_TOKEN" \
      -H "Content-Type: application/json" \
      https://api.meetzy.ai/v1/assistants
    ```
  </Tab>
</Tabs>

## Managing Existing Tokens

### Viewing Token Information

Your token list displays:

* **Token Name** - The descriptive name you assigned
* **Masked Value** - First 24 characters + masked remainder for identification
* **Expiration Date** - When the token will expire
* **Status** - Active/Expired indicator

### Copying Token Values

<Steps>
  <Step title="Locate Your Token">
    Find the token you need in your token list
  </Step>

  <Step title="Click Copy Button">
    Hover over the token row and click the "Copy" button that appears
  </Step>

  <Step title="Verify Copy Success">
    The button will briefly show "Copied!" to confirm the action
  </Step>
</Steps>

<Warning>
  You can only copy the full token value if you've just created it. For existing tokens, only the masked version is available for security.
</Warning>

## Revoking Access

When you need to revoke a token:

<Steps>
  <Step title="Identify the Token">
    Locate the token you want to revoke in your token list
  </Step>

  <Step title="Click Delete Button">
    Hover over the token row and click the trash icon
  </Step>

  <Step title="Confirm Revocation">
    Confirm the action in the popup dialog
  </Step>

  <Step title="Update Your Applications">
    Remove or replace the revoked token in any applications using it
  </Step>
</Steps>

<Info>
  Revoked tokens stop working immediately. Make sure to update any applications or services using the token before revoking it.
</Info>

## Token Expiration Handling

### Monitoring Expiration

* Check expiration dates regularly in the token manager
* Set calendar reminders before tokens expire
* Monitor your application logs for authentication errors

### Renewal Process

<Steps>
  <Step title="Create New Token">
    Generate a new token before the old one expires
  </Step>

  <Step title="Update Applications">
    Replace the old token in your applications with the new one
  </Step>

  <Step title="Test Integration">
    Verify that your applications work with the new token
  </Step>

  <Step title="Revoke Old Token">
    Once confirmed working, revoke the expired/old token
  </Step>
</Steps>

## Connecting via MCP

Meetzy exposes a **Model Context Protocol (MCP)** server that gives AI assistants like Claude, Cursor, Windsurf, and ChatGPT direct, authenticated access to your calls, WhatsApp conversations, evaluations, and analytics — all using the same API tokens described above.

<Info>
  All data exposed through the MCP server is scoped to the account that owns the token. A token from your account can only read and act on that account's assistants, calls, and conversations — there is no cross-account data access.
</Info>

### Getting the MCP Server URL

<Steps>
  <Step title="Open the MCP Section">
    Scroll to the **"Connect via MCP"** section at the bottom of the API Tokens page
  </Step>

  <Step title="Copy the Server URL">
    Click the copy icon next to the MCP server URL to copy it to your clipboard
  </Step>

  <Step title="Have a Token Ready">
    Use an existing API token, or create a new one, to authenticate your MCP connection
  </Step>
</Steps>

<Warning>
  The MCP server URL alone does not grant access. Every request must include a valid Bearer token in the `Authorization` header, exactly like any other Meetzy API call.
</Warning>

### Connecting Cursor, Claude, or Windsurf

These clients (and most desktop MCP clients) are configured through a JSON snippet that defines the server URL and headers. Add the following to your client's MCP configuration file:

```json theme={null}
{
  "mcpServers": {
    "meetzy": {
      "url": "https://mcp.meetzy.ai",
      "headers": {
        "Authorization": "Bearer YOUR_API_TOKEN"
      }
    }
  }
}
```

<Tabs>
  <Tab title="Cursor">
    Open **Settings → MCP**, click **"Add new MCP server"**, and paste the JSON snippet above into the configuration editor. Replace `YOUR_API_TOKEN` with your actual token, then save and reload Cursor.
  </Tab>

  <Tab title="Claude Desktop">
    Open your `claude_desktop_config.json` file and add the `meetzy` entry under `mcpServers`, using the snippet above. Restart Claude Desktop for the new server to appear in your tools list.
  </Tab>

  <Tab title="Windsurf">
    Go to **Settings → Cascade → MCP Servers**, click **"Add Server"**, and paste the JSON configuration. Replace the placeholder token with your real API token and reload the window.
  </Tab>
</Tabs>

<Steps>
  <Step title="Click Copy Configuration">
    Use the "Copy" button next to the JSON snippet in the token manager to copy a ready-to-use configuration
  </Step>

  <Step title="Replace the Placeholder Token">
    If you copied the snippet before creating a token, replace `YOUR_API_TOKEN` with your real token value
  </Step>

  <Step title="Paste into Your Client">
    Add the snippet to your client's MCP configuration file as described above
  </Step>

  <Step title="Restart the Client">
    Restart or reload the application so it picks up the new MCP server
  </Step>
</Steps>

### Connecting ChatGPT as a Custom Connector

ChatGPT supports MCP servers through **Custom Connectors**, authenticated with a Bearer token rather than a JSON config file.

<Steps>
  <Step title="Open Connector Settings">
    In ChatGPT, go to **Settings → Connectors → Advanced → Custom Connectors**
  </Step>

  <Step title="Add a New Connector">
    Click **"Add custom connector"** and give it a name, such as "Meetzy"
  </Step>

  <Step title="Enter the Server URL">
    Paste the MCP server URL from the token manager into the connector's URL field
  </Step>

  <Step title="Set Authentication to Bearer Token">
    Choose **Bearer Token** as the authentication method and paste your API token as the value
  </Step>

  <Step title="Save and Enable">
    Save the connector, then enable it in your conversation to let ChatGPT call Meetzy tools
  </Step>
</Steps>

<Info>
  ChatGPT sends the Bearer token in the `Authorization` header on every request, so you can reuse the same token you use for direct API calls or other MCP clients.
</Info>

### Generic Connection (Any MCP-Compatible Client)

The Meetzy MCP server implements the **Streamable HTTP** transport, so any MCP-compatible client can connect to it directly without special integration work. All you need is:

* **Server URL**: the MCP endpoint shown in the token manager
* **Transport**: Streamable HTTP
* **Authentication**: `Authorization: Bearer <your_api_token>` header on every request

<CodeGroup>
  ```bash cURL theme={null}
  curl -X POST https://mcp.meetzy.ai \
    -H "Authorization: Bearer YOUR_API_TOKEN" \
    -H "Content-Type: application/json" \
    -H "Accept: application/json, text/event-stream" \
    -d '{"jsonrpc":"2.0","method":"tools/list","id":1}'
  ```

  ```json Generic client config theme={null}
  {
    "url": "https://mcp.meetzy.ai",
    "transport": "streamable-http",
    "headers": {
      "Authorization": "Bearer YOUR_API_TOKEN"
    }
  }
  ```
</CodeGroup>

<Warning>
  Do not hardcode your API token inside shared or version-controlled MCP configuration files. Use environment variable substitution where your MCP client supports it.
</Warning>

## Troubleshooting

### Authentication Errors

If you receive 401 Unauthorized errors:

* Verify the token is correctly included in the Authorization header
* Check that the token hasn't expired
* Ensure there are no extra spaces or characters in the token value

### Token Not Working

* Confirm you're using the complete token value
* Verify the token hasn't been revoked
* Check that your account has the necessary permissions for the API endpoint

### MCP Connection Issues

* Make sure your MCP client supports the **Streamable HTTP** transport
* Confirm the `Authorization` header is being sent on every request, including the initial handshake
* If a tool call returns no data, verify the token belongs to the account that owns the assistants or calls you're trying to access

### Rate Limiting

API tokens are subject to rate limits:

* Implement exponential backoff for retries
* Monitor response headers for rate limit information
* Consider caching responses when appropriate


# Permissions & Roles
Source: https://docs.meetzy.es/account/permissions

Understanding user permissions and access control in Meetzy.

## Overview

Meetzy uses a permission-based system to control what each team member can do within the platform. Permissions are assigned individually when inviting or editing members.

<Note>
  A member can now be promoted to **Account Admin**, a role with full access to the account equivalent to the owner's, with the only exception being billing management. This role change is managed from the **Team** screen.
</Note>

## Account Admin Role

In addition to individual feature permissions, Meetzy supports an account-level role upgrade: **Account Admin**.

| Role              | Access Level                                                                |
| ----------------- | --------------------------------------------------------------------------- |
| **Owner**         | Full access to the account, including billing and subscription management   |
| **Account Admin** | Full access to the account, equivalent to the owner, **except for billing** |
| **Member**        | Access defined individually by the permissions below                        |

### How to promote a member to Account Admin

<Steps>
  <Step title="Go to the Team screen">
    Navigate to **Team** from the main menu.
  </Step>

  <Step title="Open the member's actions menu">
    Find the member you want to promote and click the actions menu (⋮) next to their name.
  </Step>

  <Step title="Edit permissions">
    Click **Edit permissions** to open the member's settings.
  </Step>

  <Step title="Set the role to Admin">
    Change the member's role to **Admin** and save changes.
  </Step>
</Steps>

<Warning>
  Account Admins can manage the team, invite or remove members, and access every area of the account (except billing). Only promote members you fully trust.
</Warning>

## Available Permissions

### Can Edit Agents

| Status     | Access Level                                           |
| ---------- | ------------------------------------------------------ |
| ✅ Enabled  | Can modify agent settings, prompts, and configurations |
| ❌ Disabled | View-only access to assigned agents                    |

Members with this permission can:

* Edit agent names and basic settings
* Modify greetings and prompts
* Configure input/output parameters
* Change voice settings

### Can Use Advanced Editor

| Status     | Access Level                         |
| ---------- | ------------------------------------ |
| ✅ Enabled  | Full access to the Advanced Editor   |
| ❌ Disabled | Limited to basic agent configuration |

Members with this permission can:

* Access all Advanced Editor panels
* Configure Knowledge Base documents
* Set up Actions & Tools (webhooks, transfers)
* Use the Playground for testing
* Access Copilot AI assistant
* Manage version history and branches
* Publish changes to production

<Tip>
  Enable "Can use advanced editor" for team members who need to make substantial changes to agent behavior.
</Tip>

### Can Access CRM

| Status     | Access Level                |
| ---------- | --------------------------- |
| ✅ Enabled  | Full access to CRM features |
| ❌ Disabled | No access to CRM menu       |

Members with this permission can:

* View and manage contacts
* Import contacts from files or Google Sheets
* Create and manage segments
* Schedule and monitor campaigns
* Export contact data

<Note>
  CRM access is only visible if your account has CRM features enabled. The administrator must have CRM access to grant it to other members.
</Note>

### Can View Billable Minutes

| Status     | Access Level                          |
| ---------- | ------------------------------------- |
| ✅ Enabled  | Can see usage and billing information |
| ❌ Disabled | No access to billing data             |

Members with this permission can:

* View call duration statistics
* See billable minutes consumption
* Access usage reports

### Can View Recordings

| Status     | Access Level                   |
| ---------- | ------------------------------ |
| ✅ Enabled  | Can listen to call recordings  |
| ❌ Disabled | Cannot access audio recordings |

Members with this permission can:

* Play call recordings from the call details
* Download recordings (if enabled)

<Warning>
  Be careful when granting recording access. Call recordings may contain sensitive customer information.
</Warning>

### Can Send WhatsApp Campaigns

| Status     | Access Level                                      |
| ---------- | ------------------------------------------------- |
| ✅ Enabled  | Can create and manage WhatsApp outbound campaigns |
| ❌ Disabled | No access to WhatsApp campaign features           |

Members with this permission can:

* Send outbound WhatsApp messages via the API
* Create and monitor WhatsApp campaigns
* View campaign status and recipient details
* Cancel in-progress campaigns

<Note>
  WhatsApp campaigns require a configured WhatsApp Business phone number on the assistant. Contact your administrator to set up the WhatsApp integration.
</Note>

## Agent-Level Access

In addition to feature permissions, you control which agents each member can access:

| Access Type     | Description                                                     |
| --------------- | --------------------------------------------------------------- |
| **Full Access** | Member can view calls and edit the agent (if permissions allow) |
| **No Access**   | Agent is hidden from the member entirely                        |

### How Agent Access Works

1. Members only see agents they're assigned to
2. Call history is filtered to show only calls from assigned agents
3. CRM data is accessible across all assigned agents
4. Reporting shows data from assigned agents only

<Note>
  Account Admins bypass agent-level restrictions and can see and manage every agent in the account.
</Note>

## Permission Combinations

Here are common permission combinations for different roles:

### Viewer (Read-Only)

* ❌ Can edit agents
* ❌ Can use advanced editor
* ❌ Can access CRM
* ❌ Can view billable minutes
* ✅ Can view recordings

Best for: Team leads who need to monitor calls without making changes.

### Agent Operator

* ✅ Can edit agents
* ❌ Can use advanced editor
* ❌ Can access CRM
* ❌ Can view billable minutes
* ✅ Can view recordings

Best for: Team members who handle basic agent configuration.

### Agent Developer

* ✅ Can edit agents
* ✅ Can use advanced editor
* ❌ Can access CRM
* ❌ Can view billable minutes
* ✅ Can view recordings

Best for: Technical team members responsible for agent development.

### Campaign Manager

* ❌ Can edit agents
* ❌ Can use advanced editor
* ✅ Can access CRM
* ❌ Can view billable minutes
* ✅ Can view recordings
* ✅ Can send WhatsApp campaigns

Best for: Marketing or sales team members managing outbound campaigns.

### Full Access

* ✅ Can edit agents
* ✅ Can use advanced editor
* ✅ Can access CRM
* ✅ Can view billable minutes
* ✅ Can view recordings
* ✅ Can send WhatsApp campaigns

Best for: Account administrators and senior team members.

<Info>
  This combination of individual permissions is different from the **Account Admin** role described above. A member with all permissions enabled still doesn't get access to the Team screen or account-wide management unless they are also promoted to Account Admin.
</Info>

## Syncing Permissions

If a member's permissions seem incorrect, administrators can sync their Firebase claims:

1. Find the member in the list
2. Click the **Sync Claims** button (only visible for advanced editor users)
3. Confirm the sync

This updates the member's authentication tokens with the latest permissions from the database.

## FAQ

<AccordionGroup>
  <Accordion title="Can I create custom roles?">
    Currently, Meetzy uses individual permissions rather than predefined roles, with the exception of the Account Admin role. You can combine permissions to create any access level you need.
  </Accordion>

  <Accordion title="What's the difference between the owner and an Account Admin?">
    An Account Admin has full access to the account, just like the owner, with one exception: they cannot manage billing or the subscription. Everything else, including team management, is available to them.
  </Accordion>

  <Accordion title="Where do I change a member's role to Account Admin?">
    Role changes are managed from the **Team** screen. Open the member's actions menu, click **Edit permissions**, and set their role to **Admin**.
  </Accordion>

  <Accordion title="Why can't I see the CRM permission toggle?">
    The CRM permission toggle is only visible if you (as the administrator) have CRM access enabled on your own account.
  </Accordion>

  <Accordion title="Do permissions apply immediately?">
    Yes, permission changes take effect immediately. Members may need to refresh their browser to see updated access.
  </Accordion>

  <Accordion title="Can a member have access to agents from multiple companies?">
    Yes, for multi-company accounts, you can assign agents from any company. Agents are organized by company in the selection list.
  </Accordion>
</AccordionGroup>


# Support Tickets
Source: https://docs.meetzy.es/account/support-tickets

View, filter, and create support tickets for your Meetzy account from the Tickets portal.

The **Tickets** page (`/tickets`) is where you track every support request tied to your account — bugs, feature requests, integration issues, and anything else our team is helping you with. It shows a live status for each ticket, a full metrics panel, resolution-time breakdowns, and a real audit trail of what happened on each ticket, including reopenings.

<Info>
  This page shows only tickets marked as **public** for your account. Internal notes and private follow-ups tracked by our team are not shown here, but nothing you need to act on is ever hidden from you.
</Info>

## Accessing the ticket list

Go to **Tickets** in the main navigation. If your account isn't yet linked to our support system, you'll see a notice instead of a ticket list — this doesn't block you from opening new tickets, it just means historical data isn't available yet. Reach out to us and we'll get it connected.

<Steps>
  <Step title="Open the Tickets page">
    Navigate to `/tickets`. The page loads your most recent ticket data automatically.
  </Step>

  <Step title="Wait for the initial load">
    While data loads, you'll see placeholder cards in the metrics area and table. This is normal and usually takes a moment.
  </Step>

  <Step title="Refresh if needed">
    Click **Actualizar** (Refresh) at any time to pull the latest ticket statuses. This is useful right after an agent replies or resolves something.
  </Step>
</Steps>

## Understanding the metrics

At the top of the page you'll find four key numbers that summarize your account's support activity:

<Tabs>
  <Tab title="Total tickets">
    The total count of tickets matching your current filters. If no filters are active, this reflects your entire history.
  </Tab>

  <Tab title="% Resolved">
    The share of tickets marked as **resolved** (`solventado`), calculated over resolvable tickets only.

    <Info>
      Discarded tickets — those closed because a doubt was clarified, the request was a duplicate, or it was out of scope — are excluded from this percentage. They aren't a support failure, so they don't count against the resolution rate.
    </Info>
  </Tab>

  <Tab title="Open now">
    Tickets that are not yet closed: pending, in progress, in testing, or in backlog.
  </Tab>

  <Tab title="Median days to close">
    The median number of days between a ticket's creation and its resolution date, based on tickets that have actually been closed.

    <Warning>
      This is a **median**, not an average. A few very old or unusually complex tickets won't skew this number the way they would an average — that's intentional, since it better reflects the typical resolution time. If the average diverges sharply from the median, the page flags it so you know the mean isn't representative.
    </Warning>
  </Tab>
</Tabs>

Below the metric cards, a summary line breaks down the exact counts per status (resolved, discarded, pending, in progress, testing, backlog). If you have any filters applied, a note reminds you that the metrics reflect the filtered subset, not your full history.

### Volume by month

Further down, a simple bar chart shows how many tickets were opened per month. The current (incomplete) month is marked with an asterisk (`*`) and rendered with a different style — don't read a lower bar there as an actual drop in volume; it just hasn't finished yet.

### Most common topics

A horizontal bar list shows which ticket types are most frequent (e.g., Bug, Feature request, Integration). Clicking a bar toggles a filter for that type, letting you drill into just that category. Click it again to remove the filter.

### Resolution times by type and priority

Below the topics list, a table breaks down resolution speed two ways: by **ticket type** and by **priority**. For each row you'll see how many tickets were closed and the median number of days it took. The type or priority with the most closed tickets is highlighted, since it's the most statistically meaningful row.

<Info>
  Only tickets with an actual resolution date are counted here — tickets still open don't have a resolution time yet, so they're excluded from both the overall median and these breakdowns.
</Info>

## Using the search and filters

The ticket list supports several filters that can be combined:

| Filter          | What it does                                                                                                                                                                                                                                                      |
| --------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Search**      | Matches against ticket title, description, ticket number, company name, and assigned agent names.                                                                                                                                                                 |
| **Mis tickets** | Shows only the tickets you personally opened from this dashboard. Tickets opened directly by the Meetzy team in our internal system aren't attributable to any specific user and won't appear under this filter, even though they're visible under "all tickets". |
| **Status**      | Filter by `pendiente`, `en_proceso`, `testing`, `solventado`, `backlog`, or `descartado`.                                                                                                                                                                         |
| **Type**        | Filter by ticket type (Bug, Performance, Prompt/LLM, UX improvement, Feature request, New agent, Integration, Data/Analytics, Other).                                                                                                                             |
| **Priority**    | Filter by criticality: Critical, High, Medium, Low.                                                                                                                                                                                                               |
| **Account**     | Only visible to Meetzy staff, who see public tickets across every account. Lets our team narrow the view down to a single customer.                                                                                                                               |
| **Date range**  | Filter by ticket creation date, using a "from" and "to" date. The "to" date is inclusive of the entire day.                                                                                                                                                       |

<Info>
  All filters apply instantly and update the metrics, monthly chart, topic breakdown, and resolution-time table to match. Use the clear filters option to reset everything back to your full ticket history.
</Info>

## Opening the ticket detail panel

Click on any ticket row to open its detail panel. This shows:

* Full title and description
* Current status, priority, urgency, and impact
* Creation, update, and (if applicable) resolution timestamps
* Who created the ticket — for tickets opened from this dashboard, you'll see the member's name or the account owner's email. Tickets opened directly by the Meetzy team don't show a creator, since they aren't tied to a specific user in your account.
* Due date, if one was set
* Resolution notes, once the ticket is closed
* Assigned agents
* Any subtasks tracked internally against the ticket
* Attached files or reference URLs
* The full **resolution history**: every time the ticket was resolved and every time it was reopened, with the date, the actor, and the note or reason attached to that event

<Warning>
  The detail panel loads ticket data on demand — if you have a slow connection, allow a moment for it to populate after clicking a row.
</Warning>

### Resolution history

Unlike the resolution notes field — which gets overwritten every time a ticket is resolved and loses any trace of prior cycles — the history section is a full, ordered timeline built from the underlying audit log. If a ticket was resolved, reopened, and resolved again, you'll see both resolution events and the reopening in between, each with its own note or reason. This is the most reliable place to understand what actually happened on a ticket that went back and forth.

## Reopening a resolved ticket

If a resolved ticket didn't actually fix your issue, you can reopen it directly from the detail panel instead of creating a duplicate.

<Steps>
  <Step title="Check the reopen window">
    Reopening is only available within **5 days** of the ticket's resolution date. After that window closes, the "Reabrir" (Reopen) option disappears and you'll need to open a new ticket referencing the old one.
  </Step>

  <Step title="Check the reopen count">
    Each ticket can be reopened a maximum of **3 times**. Once that limit is reached, the option is no longer available, even within the 5-day window.
  </Step>

  <Step title="Provide a reason">
    A reopening reason is **mandatory** — the system rejects the request without one. Be specific about what's still broken or missing; this reason becomes part of the ticket's permanent history.
  </Step>

  <Step title="Confirm">
    Once reopened, the ticket returns to an active status and the reopening is logged immediately in its resolution history, visible to both you and our team.
  </Step>
</Steps>

<Warning>
  The 5-day window and the 3-reopening cap are enforced by the backend, not just the interface — if a reopen request falls outside either limit, it will be rejected even if you find a way to trigger it manually.
</Warning>

## Comments and attachments

From the ticket detail panel you can add comments to keep the conversation going with our team. Comments support file attachments, and the same limits apply everywhere attachments are used — on ticket creation and on comments alike:

<Info>
  * Up to **10 files** per submission
  * **20 MB** maximum size per file
  * Accepted formats: images, PDF, and Excel spreadsheets (`.xls`, `.xlsx`)
</Info>

<Warning>
  You can't add comments to a ticket that's already marked as resolved or discarded — reopen it first if you need to continue the conversation.
</Warning>

## Creating a new ticket

<Steps>
  <Step title="Click 'Nuevo ticket'">
    This button is always available at the top of the page, even if your account isn't yet linked to support history.
  </Step>

  <Step title="Fill in the details">
    Provide a clear title and description. Choose a ticket type from the dropdown — this helps route your request to the right team faster.
  </Step>

  <Step title="Select the affected agent(s)">
    Choose at least one of your account's agents that the ticket relates to. This field is **mandatory** — without it, the ticket can't be submitted. If your account has several agents, this is what lets our system route the ticket to the right project internally.
  </Step>

  <Step title="Optionally, set the perceived importance">
    You can indicate how important this issue feels to you, independent of the priority our team assigns. This is purely informational and helps give context, but it doesn't override our internal triage.
  </Step>

  <Step title="Attach files if useful">
    Screenshots, logs, or spreadsheets can be attached directly (up to 10 files, 20 MB each). Attachments upload in the background after the ticket is created — a slow upload won't delay ticket creation itself.
  </Step>

  <Step title="Submit">
    Once submitted, your ticket is created and immediately visible in your list, tagged as created by you. It will also sync with our internal tracking system automatically.
  </Step>
</Steps>

<Warning>
  If your account has more than one agent and you don't select at least one, the ticket submission will be rejected. Make sure the agent selector loads correctly before submitting — if it appears empty, try refreshing the page.
</Warning>

<Info>
  If you submit a ticket while offline or the request times out, it's safe to retry — duplicate submissions are automatically prevented.
</Info>

## FAQ

<Tabs>
  <Tab title="Why don't I see all my tickets?">
    Only tickets marked as public for your account appear here. If you believe something is missing, contact your Meetzy representative.
  </Tab>

  <Tab title="Why is the % resolved lower than I expected?">
    Check whether you have status filters applied — the percentage is calculated only over the currently filtered ticket set.
  </Tab>

  <Tab title="Can I edit a ticket after creating it?">
    Ticket details can't be edited from this portal after submission. To add information, reply through the ticket detail panel or contact support directly.
  </Tab>

  <Tab title="Why can't I reopen this resolved ticket?">
    Either more than 5 days have passed since it was resolved, or it has already reached the maximum of 3 reopenings. In both cases, open a new ticket and reference the original one in the description.
  </Tab>

  <Tab title="Why doesn't 'Mis tickets' show a ticket I know exists?">
    That filter only shows tickets created from this dashboard by you. Tickets opened directly by the Meetzy team on your behalf aren't attributed to any specific user, so they won't appear here — but they're still visible under "all tickets".
  </Tab>
</Tabs>


# Team Members
Source: https://docs.meetzy.es/account/team-members

Add team members to your Meetzy account and manage their access.

## Overview

The **Members** section allows account administrators to invite team members, assign them to specific agents, promote trusted teammates to admin, and control their permissions within the platform.

## Access Team Members

1. Navigate to **Account** in the main menu
2. Select **Members**
3. View all current team members

## Team Members List

The members list displays:

| Column                   | Description                                                                            |
| ------------------------ | -------------------------------------------------------------------------------------- |
| 👤 **Avatar**            | Profile picture or initial, with a status dot (green = active, gray = invite pending)  |
| **Name**                 | Member's full name and job title                                                       |
| **Role badge**           | **Admin** or **Member**, shown next to the name                                        |
| **Invite pending badge** | Appears if the member hasn't signed in yet                                             |
| **Email**                | Member's email address                                                                 |
| **Actions**              | Edit permissions, Change email, Change password, Resend invite, Promote/Demote, Remove |

<Info>
  A gray status dot and an **Invite pending** badge next to a member's name mean they have not yet accepted their invitation or signed in for the first time. Once they log in, the dot turns green and the badge disappears.
</Info>

## Invite a New Member

Click the **+ Invite** button to add a new team member.

### Required Information

| Field            | Description                             |
| ---------------- | --------------------------------------- |
| 📧 **Email**     | The member's email address (required)   |
| 👤 **Name**      | Full name of the team member (required) |
| 💼 **Job Title** | Role or position (optional)             |

### Choose a Role

Every invitation now includes a **role selector** with two options:

| Role       | Description                                                                                                           |
| ---------- | --------------------------------------------------------------------------------------------------------------------- |
| **Member** | Standard access, limited to the permissions and agents you assign below                                               |
| **Admin**  | Full account administrator. Can manage team members, agents, and (if enabled) API tokens, just like the account owner |

<Warning>
  Only grant the **Admin** role to people you fully trust. Admins can invite, edit, promote, and remove other team members, including other admins.
</Warning>

### Permission Settings

Configure what the new member can do (these settings are ignored if the role is **Admin**, since admins have full access):

| Permission                    | Description                                                |
| ----------------------------- | ---------------------------------------------------------- |
| **Can edit agents**           | Allow modifying agent configurations                       |
| **Can use advanced editor**   | Access to the full Advanced Editor for agent configuration |
| **Can access CRM**            | Access to the CRM module (contacts, segments, campaigns)   |
| **Can view billable minutes** | See billing information and usage statistics               |
| **Can view recordings**       | Listen to call recordings                                  |

### Agent Assignment

Select which agents the member can view or edit:

* Check the boxes next to each agent
* Members will only see calls and data from assigned agents
* For multi-company accounts, agents are organized by company

### Send Invitation Email

Toggle **Send invitation email** to control how the account is created:

* **On** *(default)*: Sends a welcome email with a link to set a password and sign in
* **Off**: Creates the account silently, without sending an email — useful when you plan to share credentials directly or set a password yourself

<Info>
  The invite button label updates automatically based on this toggle: **Create & Send Invite** when the toggle is on, **Create Member** when it's off.
</Info>

Click the invite button to add the member. If the invitation email is enabled, the new teammate will show up with an **Invite pending** badge until they accept it.

## Edit Member Details

Click **Edit permissions** on any member to modify their settings:

1. Update their **Name** or **Job Title**
2. Change their **Role** (Member ↔ Admin)
3. Adjust permission toggles
4. Change agent assignments
5. Click **Save changes**

## Promote or Demote an Admin

You can change a member's role directly from the actions menu, without opening the full edit modal:

<Steps>
  <Step title="Open the actions menu">
    Click the **⋮** icon on the member's row.
  </Step>

  <Step title="Choose the action">
    * Select **Promote to admin** to grant full account administrator access
    * Select **Demote to member** to remove admin privileges and fall back to their configured permissions
  </Step>

  <Step title="Confirm">
    The role badge next to the member's name updates immediately.
  </Step>
</Steps>

<Warning>
  You cannot demote or remove yourself, and the account owner's role cannot be changed. This prevents accidentally locking yourself — or everyone else — out of the account.
</Warning>

## Resend a Pending Invitation

If a teammate hasn't accepted their invite yet (indicated by the **Invite pending** badge), you can resend it:

1. Open the **actions menu** (⋮) on the member's row
2. Click **Resend invite**
3. The button shows **Sending...** while the email is dispatched

<Info>
  This option only appears for members whose account is still pending activation.
</Info>

## Change a Teammate's Email

Administrators can update a team member's sign-in email directly from the member list:

1. Open the **actions menu** (⋮) on the member's row
2. Click **Change email**
3. Enter the new email address
4. Confirm the change

<Note>
  This updates where the member signs in and receives notifications. The member may need to use the new address the next time they log in.
</Note>

## Change Member Password

Administrators can reset a member's password:

1. Click **Change Password** on the member
2. Enter the new password (minimum 6 characters)
3. Click **Update Password**

The member will need to use the new password on their next login.

## Remove a Member

To remove a team member:

1. Click **Remove** on the member's row
2. Confirm the removal in the dialog

<Warning>
  Removing a member is permanent and cannot be undone. All their data and access will be permanently deleted.
</Warning>

## Search Members

Use the search bar to find members by name:

* Type in the search field
* Results filter as you type

## Best Practices

1. **Principle of least privilege**: Only grant the **Admin** role or elevated permissions when a teammate truly needs them
2. **Agent-specific access**: Assign members only to agents they work with
3. **Watch pending invites**: Follow up or resend invitations that stay pending for too long
4. **Regular audits**: Periodically review member roles, permissions, and remove inactive users
5. **Secure passwords**: Use strong passwords when creating accounts without email invitations


# Actions & Tools
Source: https://docs.meetzy.es/advanced-editor/actions-tools

Configure integrations, call transfers, webhooks, and automated actions for your AI agents

## Overview

The Actions & Tools section is the powerhouse of your AI agent's operational capabilities. Here you configure what your agent can do during and after calls - from transferring conversations to human agents or SIP systems, to making intelligent API calls, checking business hours, detecting voicemail early, and setting up sophisticated callback protocols.

<Info>
  This section transforms your agent from a simple conversationalist into a fully integrated business automation tool that can take real actions and connect with your existing systems.
</Info>

## Call Transfer

Configure how your agent intelligently transfers calls to human agents, departments, or external phone systems when needed.

### Transfer Configuration

<Steps>
  <Step title="Enable Transfer Capability">
    Toggle the "Call Transfer" option to activate the transfer functionality for your agent.
  </Step>

  <Step title="Choose Transfer Destination">
    Select between traditional phone numbers or SIP trunk endpoints based on your infrastructure.
  </Step>

  <Step title="Configure Transfer Logic">
    Define when and how your agent should initiate transfers using smart prompting and multi-language templates.
  </Step>

  <Step title="Test Transfer Flow">
    Validate that transfers work correctly with your phone system before going live.
  </Step>
</Steps>

### Transfer Types

<Tabs>
  <Tab title="Phone Number">
    **Traditional Phone Transfer**

    Perfect for transferring to mobile phones, landlines, or traditional business numbers.

    * **Format**: Use international format (+1234567890)
    * **Best for**: Human agents, customer service departments, sales teams
    * **Reliability**: High compatibility with all phone systems
    * **Cost**: Standard telephony rates apply

    **Configuration Example:**

    ```
    Transfer Number: +1-555-123-4567
    Department: Sales Team
    ```
  </Tab>

  <Tab title="SIP Trunk">
    **SIP/VoIP Transfer**

    Ideal for enterprise environments with PBX systems, call centers, or VoIP infrastructure.

    * **Format**: `sip:extension@gateway_ip:port`
    * **Best for**: Call centers, PBX extensions, enterprise VoIP
    * **Benefits**: Lower costs, advanced routing, better integration
    * **Requirements**: SIP-compatible phone system

    **Configuration Examples:**

    ```
    Sales Extension: sip:100@10.0.1.50:5060
    Support Queue: sip:support@pbx.company.com:5060  
    Manager Direct: sip:200@192.168.1.100:5060
    ```
  </Tab>
</Tabs>

### Transfer Execution Mode

Beyond choosing *where* a call goes, you can now control exactly *when* the transfer is executed relative to what the agent says right before handing off the call. This setting applies to the transfer tool as a whole, regardless of whether you're using the classic single-destination configuration or Transfer v2 rules.

<Tabs>
  <Tab title="Post Tool Speech (Default)">
    **Speak First, Then Transfer**

    The agent finishes saying the transfer message — the default phrase or the rule-specific Customer Message — and only then does the actual call transfer take place.

    * **Behavior**: The caller always hears a clear heads-up ("Let me connect you now\...") before the line is handed off
    * **Best for**: Most use cases, especially Warm transfers where a smooth, predictable handoff matters
    * **Trade-off**: A short, natural pause between the message and the actual transfer
    * **Recommended**: This is the default behavior for the Call Transfer tool
  </Tab>

  <Tab title="Immediate">
    **Transfer Right Away**

    The transfer is triggered as soon as the agent decides to transfer, without waiting for the message to finish playing — the handoff and the spoken message effectively overlap.

    * **Behavior**: Fastest possible handoff, but the caller may hear the message cut short as the transfer connects
    * **Best for**: Time-sensitive scenarios where shaving off every second matters more than a polished announcement
    * **Trade-off**: Less predictable caller experience compared to Post Tool Speech
  </Tab>
</Tabs>

<Info>
  If you don't explicitly configure an execution mode, your agent automatically uses **Post Tool Speech** — this ensures the caller always hears the transfer message before being handed off.
</Info>

### Transfer v2 (Beta): Multiple Transfer Rules

<Info>
  **Transfer v2** is a Beta upgrade to Call Transfer that replaces the single destination model with a flexible list of **transfer rules**. Instead of sending every transfer to the same number, your agent evaluates each rule's condition in real time and routes the call to the matching destination — with its own type, customer message, and optional briefing for the receiving human agent.
</Info>

Enable Transfer v2 with the **"Transfer to multiple destinations (Beta)"** toggle inside the Call Transfer panel. Once active, you can define as many rules as you need.

Each rule includes:

| Field                 | Description                                                                                                                                                    |
| --------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Condition**         | A natural-language description of when this rule should trigger (e.g., "the customer asks to speak with sales", "the customer wants to cancel a subscription") |
| **Destination Type**  | `Phone` or `SIP`, matching the same formats used in standard Call Transfer                                                                                     |
| **Destination Value** | The phone number (`+1234567890`) or SIP URI (`sip:extension@ip:port`) to route to                                                                              |
| **Transfer Type**     | `Direct` or `Warm` — controls how the handoff is performed                                                                                                     |
| **Customer Message**  | An optional per-rule message the agent says to the caller right before transferring, overriding the default transfer phrase                                    |
| **Agent Briefing**    | An optional message relayed to the receiving human agent — a short summary of who is calling and why — used only on `Warm` transfers                           |

<Tabs>
  <Tab title="Direct Transfer">
    **Direct (Blind / Cold) Transfer**

    The agent immediately connects the caller to the destination without staying on the line.

    * **Best for**: Simple routing to a known extension or department
    * **Experience**: Fastest handoff, no context relay
    * **Use when**: The destination doesn't need a live briefing before speaking with the customer
    * **Agent Briefing**: Not applicable — since the agent disconnects immediately, there's no opportunity to relay context
  </Tab>

  <Tab title="Warm Transfer">
    **Warm (Conference) Transfer**

    The agent stays connected momentarily to introduce the call and pass context before disconnecting.

    * **Best for**: Complex cases, VIP customers, or situations requiring a smooth handoff
    * **Experience**: The receiving human agent hears the Agent Briefing before taking over the conversation
    * **Use when**: Context preservation is critical to avoid the customer repeating themselves
    * **Agent Briefing**: Recommended — write a short, factual summary such as "Caller is asking about invoice #4521, already confirmed their identity"
  </Tab>
</Tabs>

<Steps>
  <Step title="Enable Transfer v2">
    Toggle "Transfer to multiple destinations (Beta)" within the Call Transfer section.
  </Step>

  <Step title="Add a Rule">
    Click "Add rule" and describe the condition that should trigger this specific transfer.
  </Step>

  <Step title="Configure Destination">
    Choose Phone or SIP and enter the corresponding destination value.
  </Step>

  <Step title="Choose Transfer Type">
    Select Direct for an immediate handoff or Warm to relay context before disconnecting.
  </Step>

  <Step title="Add a Customer Message (Optional)">
    Write a message specific to this rule, such as "Let me connect you with our billing team" for a billing-related rule.
  </Step>

  <Step title="Add an Agent Briefing (Optional, Warm only)">
    For Warm transfers, write a short internal summary the receiving human agent will hear before taking the call — for example, the customer's issue, sentiment, or anything already resolved.
  </Step>

  <Step title="Repeat for Additional Rules">
    Add as many rules as needed to cover all the transfer scenarios for your agent.
  </Step>
</Steps>

<Warning>
  Rules are evaluated in the order they appear. Place more specific conditions above general ones to avoid a broad rule capturing calls meant for a more precise destination.
</Warning>

**Example Configuration:**

```
Rule 1
Condition: the customer wants to talk to sales or ask about pricing
Destination: Phone → +1-555-200-1000
Type: Warm
Customer Message: "I'll connect you with our sales team, one moment please."
Agent Briefing: "Inbound caller interested in the Pro plan, already asked about annual pricing."

Rule 2
Condition: the customer reports a technical issue or outage
Destination: SIP → sip:support@pbx.company.com:5060
Type: Direct
Customer Message: "Transferring you to technical support now."
Agent Briefing: (not applicable — Direct transfer)

Rule 3
Condition: the customer explicitly asks for a manager or wants to file a complaint
Destination: Phone → +1-555-200-2000
Type: Warm
Customer Message: (default transfer phrase)
Agent Briefing: "Caller is upset about a delayed refund, already apologized and confirmed order #8823."
```

<Info>
  When Transfer v2 is disabled, your agent falls back to the classic single-destination Call Transfer configuration described above. The Transfer Execution Mode setting applies the same way in both cases.
</Info>

<Warning>
  Transfer v2 is currently in **Beta**. Test each rule's condition thoroughly with realistic phrasing before relying on it in production, and keep an eye on how overlapping conditions are prioritized.
</Warning>

### Multi-Language Transfer Prompt Templates

The transfer prompt is crucial - it teaches your AI agent exactly when to transfer calls. Choose from professionally crafted templates in 6 languages:

<Tabs>
  <Tab title="Spanish">
    ```
    Se usa para transferir la llamada con una persona en los casos que se indican a continuación:

    - El usuario rechaza directamente la propuesta o servicio
    - El usuario solicita expresamente hablar con una persona humana
    - El usuario muestra desconfianza hacia el asistente virtual
    - Se necesita información específica que no está disponible en el sistema
    - El usuario requiere autorización o gestión de datos personales sensibles
    - Situaciones que requieren negociación o decisiones complejas
    - Quejas o reclamaciones que necesitan atención personalizada
    - El usuario solicita escalación a un supervisor o gerente
    - Casos de facturación complejos o disputas de pagos
    - Situaciones médicas o legales que requieren experiencia humana
    ```
  </Tab>

  <Tab title="English">
    ```
    Used to transfer the call to a person in the following cases:

    - The user directly rejects the proposal or service
    - The user expressly requests to speak with a human person
    - The user shows distrust towards the virtual assistant
    - Specific information is needed that is not available in the system
    - The user requires authorization or management of sensitive personal data
    - Situations requiring negotiation or complex decisions
    - Complaints or claims that need personalized attention
    - The user requests escalation to a supervisor or manager
    - Complex billing or payment dispute situations
    - Medical or legal situations requiring human expertise
    ```
  </Tab>

  <Tab title="French">
    ```
    Utilisé pour transférer l'appel à une personne dans les cas suivants:

    - L'utilisateur rejette directement la proposition ou le service
    - L'utilisateur demande expressément à parler avec une personne humaine
    - L'utilisateur montre de la méfiance envers l'assistant virtuel
    - Des informations spécifiques sont nécessaires qui ne sont pas disponibles dans le système
    - L'utilisateur nécessite une autorisation ou la gestion de données personnelles sensibles
    - Situations nécessitant négociation ou décisions complexes
    - Plaintes ou réclamations nécessitant une attention personnalisée
    - L'utilisateur demande une escalade vers un superviseur ou manager
    - Situations de facturation complexes ou litiges de paiement
    - Situations médicales ou légales nécessitant une expertise humaine
    ```
  </Tab>

  <Tab title="German">
    ```
    Wird verwendet, um den Anruf an eine Person zu übertragen in folgenden Fällen:

    - Der Benutzer das Angebot oder den Service direkt ablehnt
    - Der Benutzer ausdrücklich darum bittet, mit einer menschlichen Person zu sprechen
    - Der Benutzer Misstrauen gegenüber dem virtuellen Assistenten zeigt
    - Spezifische Informationen benötigt werden, die im System nicht verfügbar sind
    - Der Benutzer eine Autorisierung oder Verwaltung sensibler persönlicher Daten benötigt
    - Situationen, die Verhandlungen oder komplexe Entscheidungen erfordern
    - Beschwerden oder Ansprüche, die persönliche Aufmerksamkeit benötigen
    - Der Benutzer eine Eskalation zu einem Vorgesetzten oder Manager anfordert
    - Komplexe Abrechnungs- oder Zahlungsstreitigkeiten
    - Medizinische oder rechtliche Situationen, die menschliche Expertise erfordern
    ```
  </Tab>

  <Tab title="Italian">
    ```
    Usato per trasferire la chiamata a una persona nei seguenti casi:

    - L'utente rifiuta direttamente la proposta o il servizio
    - L'utente richiede espressamente di parlare con una persona umana
    - L'utente mostra sfiducia verso l'assistente virtuale
    - Sono necessarie informazioni specifiche non disponibili nel sistema
    - L'utente richiede autorizzazione o gestione di dati personali sensibili
    - Situazioni che richiedono negoziazione o decisioni complesse
    - Reclami o richieste che necessitano attenzione personalizzata
    - L'utente richiede escalation a un supervisore o manager
    - Situazioni di fatturazione complesse o dispute di pagamento
    - Situazioni mediche o legali che richiedono competenza umana
    ```
  </Tab>

  <Tab title="Portuguese">
    ```
    Usado para transferir a chamada para uma pessoa nos seguintes casos:

    - O usuário rejeita diretamente a proposta ou serviço
    - O usuário solicita expressamente falar com uma pessoa humana
    - O usuário mostra desconfiança em relação ao assistente virtual
    - São necessárias informações específicas que não estão disponíveis no sistema
    - O usuário requer autorização ou gestão de dados pessoais sensíveis
    - Situações que requerem negociação ou decisões complexas
    - Reclamações ou solicitações que precisam de atenção personalizada
    - O usuário solicita escalação para um supervisor ou gerente
    - Situações complexas de faturamento ou disputas de pagamento
    - Situações médicas ou legais que requerem expertise humana
    ```
  </Tab>
</Tabs>

## Webhook Tools

Create powerful API integrations that allow your agent to interact with external systems during conversations through our sophisticated modal interface.

<Info>
  **Enhanced Modal Interface**: Webhook Tools now feature an advanced configuration modal with real-time testing, parameter mapping, authentication methods, and intelligent conversation data extraction.
</Info>

### Webhook Tool Modal Interface

Configure webhook tools through our comprehensive modal interface:

<Steps>
  <Step title="Tool Identity Configuration">
    Set up the basic tool identity and activation conditions.

    **Required Fields:**

    * **Name**: Unique identifier (e.g., "create\_lead", "check\_inventory")
    * **Description**: Clear explanation of when the AI should use this tool
    * **Execution Mode**: Choose immediate or deferred execution
  </Step>

  <Step title="API Endpoint Setup">
    Configure the core API connection details.

    **Configuration Options:**

    * **Base URL**: Complete webhook endpoint URL
    * **HTTP Method**: GET, POST, PUT, PATCH, or DELETE
    * **Request Timeout**: 5-60 seconds (default: 20)
    * **Error Handling**: Retry logic and fallback behavior
  </Step>

  <Step title="Authentication Configuration">
    Set up secure authentication for your API endpoints.

    **Supported Authentication Types:**

    * Bearer Token authentication (OAuth 2.0, JWT)
    * API Key authentication (Header or Query parameter)
    * Basic Auth with username/password
    * Custom header authentication
  </Step>

  <Step title="Parameter Mapping">
    Define intelligent parameter mapping with conversation data extraction.

    **Parameter Configuration:**

    * **Query Parameters**: URL-encoded for GET requests
    * **Body Parameters**: JSON payload for POST/PUT/PATCH
    * **Header Parameters**: Custom HTTP headers
    * **Value Sources**: LLM extraction, static values, conversation metadata
  </Step>
</Steps>

### Advanced Parameter Configuration

Each webhook parameter supports sophisticated data mapping:

<Tabs>
  <Tab title="Parameter Properties">
    | Property              | Description              | Options                                         | Example                                  |
    | --------------------- | ------------------------ | ----------------------------------------------- | ---------------------------------------- |
    | **Name**              | Parameter identifier     | Custom string                                   | `customer_name`                          |
    | **Type**              | Data validation          | string, number, boolean, integer, array, object | `string`                                 |
    | **Required**          | Mandatory field          | true/false                                      | `true`                                   |
    | **Description**       | AI understanding context | Natural language                                | "Customer's full name from conversation" |
    | **Value Source**      | Data extraction method   | LLM extraction, static value, metadata          | LLM prompt                               |
    | **Extraction Prompt** | AI instruction           | Natural language                                | "Extract the customer's email address"   |
  </Tab>

  <Tab title="Value Sources">
    **LLM Extraction**

    Intelligent conversation data extraction:

    ```json theme={null}
    {
      "extraction_prompt": "Extract the customer's phone number from the conversation",
      "parameter_type": "string",
      "required": true
    }
    ```

    **Static Values**

    Fixed values for consistent data:

    ```json theme={null}
    {
      "static_value": "phone_call",
      "parameter": "source",
      "type": "string"
    }
    ```

    **Conversation Metadata**

    Built-in conversation data:

    * Call duration and timestamps
    * Customer and agent phone numbers
    * Call direction (inbound/outbound)
    * Call status and outcome
    * Agent ID and configuration details
  </Tab>

  <Tab title="Data Types">
    **Supported Parameter Types:**

    * **string**: Text data, names, descriptions, addresses
    * **number**: Numeric values with decimal precision
    * **integer**: Whole numbers only (IDs, quantities)
    * **boolean**: true/false values for flags
    * **array**: List of values or complex objects
    * **object**: Nested data structures with properties

    **Type Validation:**
    Automatic validation ensures data consistency and provides clear error messages for type mismatches.
  </Tab>
</Tabs>

### Webhook Authentication Methods

<Tabs>
  <Tab title="Bearer Token">
    **JWT/OAuth 2.0 Authentication**

    Modern token-based authentication for secure API access:

    ```json theme={null}
    {
      "Authorization": "Bearer eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9..."
    }
    ```

    **Use Cases:**

    * Salesforce, HubSpot, and modern CRM systems
    * OAuth 2.0 protected endpoints
    * JWT-based authentication systems
    * Enterprise SaaS platforms
  </Tab>

  <Tab title="API Key">
    **Simple API Key Authentication**

    Traditional API key authentication:

    ```json theme={null}
    {
      "X-API-Key": "sk_live_abcd1234efgh5678ijkl",
      "Authorization": "ApiKey sk_live_abcd1234efgh5678ijkl"
    }
    ```

    **Use Cases:**

    * Third-party service APIs
    * Legacy system integration
    * Custom internal APIs
    * Simple authentication requirements
  </Tab>

  <Tab title="Basic Auth">
    **Username/Password Authentication**

    HTTP Basic authentication:

    ```json theme={null}
    {
      "Authorization": "Basic dXNlcm5hbWU6cGFzc3dvcmQ="
    }
    ```

    **Use Cases:**

    * Legacy systems
    * Internal APIs
    * Simple HTTP authentication
    * Development and testing environments
  </Tab>
</Tabs>

### Webhook Integration Examples

<Tabs>
  <Tab title="CRM Lead Creation">
    **Salesforce Lead Integration**

    Automatically create leads when customers show purchase interest:

    ```json theme={null}
    {
      "name": "create_salesforce_lead",
      "description": "Create a new lead in Salesforce when customer shows purchase interest",
      "url": "https://api.salesforce.com/services/data/v52.0/sobjects/Lead",
      "method": "POST",
      "authentication": {
        "type": "bearer_token",
        "token": "{{salesforce_access_token}}"
      },
      "body_parameters": {
        "FirstName": {
          "type": "string",
          "required": true,
          "value_type": "llm_extraction",
          "extraction_prompt": "Extract customer's first name from conversation"
        },
        "LastName": {
          "type": "string",
          "required": true,
          "value_type": "llm_extraction",
          "extraction_prompt": "Extract customer's last name from conversation"
        },
        "Phone": {
          "type": "string",
          "required": true,
          "value_type": "conversation_metadata",
          "metadata_field": "customer_phone"
        },
        "LeadSource": {
          "type": "string",
          "value_type": "static",
          "static_value": "AI Phone Call"
        }
      }
    }
    ```
  </Tab>

  <Tab title="Inventory Checking">
    **Real-time Product Availability**

    Check product availability during customer inquiries:

    ```json theme={null}
    {
      "name": "check_product_availability",
      "description": "Check real-time product availability when customer asks about specific items",
      "url": "https://inventory.yourcompany.com/api/check",
      "method": "GET",
      "authentication": {
        "type": "api_key",
        "key": "{{inventory_api_key}}",
        "location": "header"
      },
      "query_parameters": {
        "product_sku": {
          "type": "string",
          "required": true,
          "value_type": "llm_extraction",
          "extraction_prompt": "Extract product SKU or name mentioned by customer"
        },
        "location": {
          "type": "string",
          "value_type": "llm_extraction",
          "extraction_prompt": "Extract customer's preferred store location or zip code"
        },
        "quantity": {
          "type": "integer",
          "value_type": "llm_extraction",
          "extraction_prompt": "Extract desired quantity, default to 1 if not specified"
        }
      },
      "response_timeout": 10
    }
    ```
  </Tab>

  <Tab title="Calendar Booking">
    **Appointment Scheduling**

    Automatically schedule appointments through calendar systems:

    ```json theme={null}
    {
      "name": "book_calendar_appointment",
      "description": "Schedule an appointment when customer requests a meeting or consultation",
      "url": "https://api.calendly.com/scheduled_events",
      "method": "POST",
      "authentication": {
        "type": "bearer_token",
        "token": "{{calendly_access_token}}"
      },
      "body_parameters": {
        "event_type": {
          "type": "string",
          "value_type": "static",
          "static_value": "https://calendly.com/your-company/30-min-consultation"
        },
        "start_time": {
          "type": "string",
          "value_type": "llm_extraction",
          "extraction_prompt": "Extract preferred meeting time from customer in ISO format"
        },
        "invitee": {
          "type": "object",
          "properties": {
            "name": {
              "type": "string",
              "value_type": "llm_extraction",
              "extraction_prompt": "Customer's full name"
            },
            "email": {
              "type": "string",
              "value_type": "llm_extraction",
              "extraction_prompt": "Customer's email address"
            }
          }
        }
      }
    }
    ```
  </Tab>
</Tabs>

## Business Hours (Beta)

<Info>
  **Business Hours** is a new built-in tool (Beta) that lets your agent automatically resolve whether the current call is happening within your configured business hours — no webhook or external calendar needed. It replaces manual date/time reasoning with a structured weekly schedule, holiday calendar, and time zone awareness.
</Info>

Instead of prompting the model to reason about dates, weekdays, and time zones, the Business Hours tool evaluates the call timestamp against your configuration and returns whether the business is **open** or **closed** at that exact moment — including support for holiday exceptions and dedicated messaging for each state.

### What It Solves

* Avoids transferring calls to a human team that isn't available
* Lets the agent give accurate opening hours or a "we're currently closed" response automatically
* Enables conditional logic such as "only offer live transfer during business hours, otherwise offer a callback"
* Accounts for holidays and special closures without extra prompt engineering
* Removes the need for manual date/time arithmetic in the system prompt

### Configuration

<Steps>
  <Step title="Enable Business Hours">
    Add the Business Hours tool from the Actions & Tools panel and toggle it on.
  </Step>

  <Step title="Set Your Weekly Schedule">
    Configure opening and closing times independently for each day of the week (Monday through Sunday). Mark any day as fully closed when it doesn't apply — for example, weekends or days your business doesn't operate.
  </Step>

  <Step title="Add Holidays">
    Define specific dates — national holidays, company closures, or one-off exceptions — that should always be treated as closed (or with custom hours) regardless of the regular weekly schedule.
  </Step>

  <Step title="Configure Time Zone">
    Select the time zone your business hours are defined in, so the tool correctly resolves the call time regardless of where the customer is calling from.
  </Step>

  <Step title="Set In-Hours and Out-of-Hours Messages">
    Write a custom message the agent uses when the call falls within business hours, and a different message for when it falls outside business hours (including holidays).
  </Step>

  <Step title="Reference the Tool in Your Prompt or Rules">
    Use the tool's result in your system prompt logic, or combine it with a Transfer v2 rule condition (e.g., "only transfer to sales if within business hours").
  </Step>
</Steps>

### Weekly Schedule Fields

Each day of the week supports its own independent configuration:

| Field            | Description                             |
| ---------------- | --------------------------------------- |
| **Day**          | Monday through Sunday                   |
| **Status**       | Open or Closed for that day             |
| **Opening Time** | Start of business hours (e.g., `09:00`) |
| **Closing Time** | End of business hours (e.g., `18:00`)   |

### Holidays

Holidays override the regular weekly schedule for a specific date:

| Field      | Description                                                    |
| ---------- | -------------------------------------------------------------- |
| **Date**   | The specific calendar date the holiday applies to              |
| **Label**  | A descriptive name (e.g., "New Year's Day", "Company Retreat") |
| **Status** | Closed all day, or Open with custom hours for that date        |

### In-Hours vs Out-of-Hours Messages

<Tabs>
  <Tab title="In-Hours Message">
    **Used while the business is open**

    Optional message the agent can use to confirm availability, such as:

    ```
    "We're currently open — I can connect you with our team right away."
    ```
  </Tab>

  <Tab title="Out-of-Hours Message">
    **Used while the business is closed (including holidays)**

    Message explaining unavailability and next steps, such as:

    ```
    "We're currently closed. Our business hours are Monday to Friday, 9 AM to 6 PM. Would you like me to schedule a callback for the next business day?"
    ```
  </Tab>
</Tabs>

### Example Schedule

```
Monday - Friday: 09:00 - 18:00
Saturday: 10:00 - 14:00
Sunday: Closed
Time Zone: Europe/Madrid

Holidays:
- Jan 1: New Year's Day (Closed)
- Dec 25: Christmas Day (Closed)
- Aug 15: Summer Break (Closed)

In-Hours Message: "We're currently open — I can connect you with our team right away."
Out-of-Hours Message: "We're currently closed. Our business hours are Monday to Friday, 9 AM to 6 PM. Would you like me to schedule a callback?"
```

### Example Usage

**Combined with Transfer v2:**

```
Rule 1
Condition: the customer wants to talk to sales AND business hours are open
Destination: Phone → +1-555-200-1000
Type: Warm

Rule 2
Condition: the customer wants to talk to sales AND business hours are closed
Destination: (no transfer — offer to schedule a callback instead)
```

**Combined with End Call Strategy or system prompt:**

Instruct the agent to check the Business Hours result before offering a live transfer, and fall back to taking a message or booking a callback when the result is `closed`.

<Warning>
  Business Hours is currently in **Beta**. Double-check the resolved status during testing, especially around daylight saving time transitions and newly added holidays.
</Warning>

## Early Voicemail Detection

<Info>
  **Early Voicemail Detection** adds a telephony-aware layer of Answering Machine Detection (AMD) that works alongside — and independently from — the phrase-based detection configured in **End Call Strategy** below. Instead of relying only on the agent recognizing a voicemail greeting mid-conversation, this feature can flag an answering machine earlier, using either the agent's own reasoning or signals from the carrier.
</Info>

Enable this tool to choose **how early** and **at what level** voicemail detection happens, trading off a small amount of latency for a higher confidence detection, or vice versa. Detection is organized into two levels — **Agent** and **Carrier** (also referred to as "Operator") — with the Carrier level offering two submodes, **Gate** and **Live**.

### Detection Levels

<Tabs>
  <Tab title="Agent">
    **Agent-Level Detection (Native)**

    The AI agent listens to the call audio itself and recognizes voicemail or answering machine greetings using its own conversational reasoning — the same underlying mechanism used by the End Call Strategy phrase templates.

    * **How it works**: The agent starts speaking normally and identifies typical voicemail phrasing as the conversation unfolds ("please leave a message after the tone", etc.)
    * **Best for**: Simple setups where a natural conversation start matters more than shaving off detection time, or when your telephony carrier doesn't provide a reliable native AMD signal
    * **When to use it**: Choose Agent when you'd rather keep the call flow simple and let the same phrase-recognition logic you already trust (via End Call Strategy) handle voicemail cases, without adding a telephony-level dependency
    * **Default**: This is the default level the first time Early Voicemail Detection is enabled
  </Tab>

  <Tab title="Carrier (Operator)">
    **Carrier-Level Detection**

    Detection is delegated to the telephony carrier's native Answering Machine Detection signal, which analyzes call setup characteristics — ringing pattern, greeting cadence, energy levels — independently of what the agent says.

    * **How it works**: The carrier analyzes the call in real time and reports whether it believes a human or a machine answered
    * **Best for**: Faster, more consistent detection, especially useful for outbound calling at scale
    * **When to use it**: Choose Carrier when you're running outbound campaigns and want detection decisions grounded in telephony-level signals rather than solely on what the agent hears and interprets
    * **Requires**: Choosing one of the two Carrier submodes below (Gate or Live)
  </Tab>
</Tabs>

### Carrier Submodes: Gate vs Live

When the Carrier level is selected, choose how the agent should act on the carrier's AMD signal:

<Tabs>
  <Tab title="Gate">
    **Gate Mode**

    The system waits for the carrier's AMD analysis to finish **before** letting the agent start speaking.

    * **Behavior**: A brief pause occurs at the very start of the call while the carrier determines human vs. machine
    * **Trade-off**: A slightly slower start, but the agent never delivers the wrong opening line to a machine
    * **When to use it**: Choose Gate for outbound calling where avoiding a wasted or awkward greeting to voicemail outweighs a one-to-two second delay at the start of the call — ideal for high-volume campaigns where consistency matters more than speed
  </Tab>

  <Tab title="Live">
    **Live Mode**

    The call connects immediately and the carrier's AMD analysis runs live, in parallel with the conversation.

    * **Behavior**: The agent can start speaking right away; if the carrier reports a machine mid-stream, the system reacts on the fly (e.g., handing off to End Call Strategy or leaving a pre-recorded message)
    * **Trade-off**: A faster, more natural-feeling start, with a small chance the agent speaks a line or two before the machine is confirmed
    * **When to use it**: Choose Live when a fast, natural call start is the priority — for example, inbound-style experiences or campaigns where caller perception of responsiveness matters more than avoiding the rare case of the agent speaking briefly to a machine
  </Tab>
</Tabs>

<Info>
  The system remembers which Carrier submode (Gate or Live) you last used. Switching back to Carrier after temporarily trying Agent level automatically restores your previous submode.
</Info>

### Configuration

<Steps>
  <Step title="Enable Early Voicemail Detection">
    Toggle on Early Voicemail Detection in the Actions & Tools panel.
  </Step>

  <Step title="Choose a Detection Level">
    Select **Agent** for native recognition by the AI, or **Carrier** (Operator) for telephony-level AMD.
  </Step>

  <Step title="Select a Submode (Carrier only)">
    If you chose Carrier, pick **Gate** for a safer, slightly delayed start, or **Live** for an immediate start with parallel analysis.
  </Step>

  <Step title="Keep End Call Strategy Active as a Fallback">
    Leave your End Call Strategy phrase templates configured — Early Voicemail Detection catches most cases sooner, while phrase detection remains a reliable safety net for anything it misses.
  </Step>
</Steps>

<Warning>
  Carrier-level AMD accuracy can vary depending on your telephony provider and the destination country. Test both Gate and Live submodes with real call patterns before relying on them for high-volume outbound campaigns.
</Warning>

## End Call Strategy

Configure intelligent call termination with sophisticated answering machine detection and strategic conversation ending across multiple languages.

### Enhanced Call Termination Logic

<Steps>
  <Step title="Enable Smart Termination">
    Activate the "End Call Strategy" to handle call endings intelligently with multi-language support.
  </Step>

  <Step title="Configure Detection Phrases">
    Use comprehensive phrase detection across 6 languages for answering machines and voicemail systems.
  </Step>

  <Step title="Define Completion Criteria">
    Specify what constitutes a successful call that should end naturally.
  </Step>

  <Step title="Handle Edge Cases">
    Configure responses for unexpected scenarios like long silences or system errors.
  </Step>
</Steps>


# Basic Settings
Source: https://docs.meetzy.es/advanced-editor/basic-settings

Configure agent name, language, voice, AI model, and foundational components with advanced features

## Overview

Basic Settings is the foundation of your AI agent configuration in Meetzy's Advanced Editor. This section establishes the core identity and capabilities of your agent, including specialized skills, professional voice selection, advanced AI model options, and powerful foundational components that ensure professional-grade performance from the first call.

## Agent Identification

### Agent Name

Define your agent's internal name for team identification and management:

<Info>
  **Naming Best Practices**: Use descriptive names like "Customer Support - English" or "Real Estate Sales - Spanish" to easily distinguish agents in your dashboard.
</Info>

### Company Selection (Admin Only)

Administrators can assign agents to specific organizations within the system:

<Warning>
  Company assignment determines available resources, phone numbers, and access permissions for the agent.
</Warning>

## Twilio Account Management (Admin Only)

<Tabs>
  <Tab title="Account Selection">
    Choose the appropriate Twilio account for new agent creation:

    <Steps>
      <Step title="Select Twilio Account">
        Choose from available Twilio accounts in the dropdown menu
      </Step>

      <Step title="Automatic Resource Loading">
        Phone numbers and resources automatically populate for the selected account
      </Step>

      <Step title="Configuration Integration">
        Agent inherits the selected account's settings and capabilities
      </Step>
    </Steps>

    <Info>
      Multiple Twilio accounts enable organization-specific configurations and billing separation.
    </Info>
  </Tab>

  <Tab title="Account Switching">
    <Warning>
      **Resource Impact**: Changing Twilio accounts reloads available phone numbers and may affect existing call routing configurations.
    </Warning>

    **Account Change Effects:**

    * Phone number availability updates
    * Caller ID verification requirements may change
    * Regional settings and capabilities may vary
  </Tab>
</Tabs>

## Advanced Phone Configuration

### Phone Number Assignment

<Tabs>
  <Tab title="New Agents">
    Configure phone numbers for new agents with comprehensive management:

    <Steps>
      <Step title="Select Twilio Account">
        Choose the appropriate Twilio account (admin users only)
      </Step>

      <Step title="Choose Available Number">
        Select from dynamically loaded available phone numbers
      </Step>

      <Step title="Request Additional Numbers">
        Use the direct Twilio link to purchase more numbers if needed
      </Step>

      <Step title="Verify Assignment">
        Confirm the number appears correctly in the configuration
      </Step>
    </Steps>

    <Info>
      **Real-time Availability**: Phone numbers are filtered by selected Twilio account and show only unassigned numbers.
    </Info>
  </Tab>

  <Tab title="Existing Agents">
    Phone numbers for existing agents are permanent and displayed for reference:

    <Warning>
      **Permanent Assignment**: Phone numbers cannot be changed after agent creation. Plan your number allocation carefully.
    </Warning>

    **Read-only Display:**

    * Shows assigned phone number in friendly format
    * Displays associated Twilio account information
    * Provides reference for configuration management
  </Tab>
</Tabs>

### Enhanced Caller ID Management

Configure and manage client phone numbers for outbound calling with advanced verification and status tracking:

<Tabs>
  <Tab title="Adding Caller IDs">
    <Steps>
      <Step title="Initiate Addition">
        Click "Add Caller ID" to open the configuration dialog
      </Step>

      <Step title="Enter Phone Information">
        Provide a descriptive name and the phone number for verification
      </Step>

      <Step title="Automatic Verification Call">
        Twilio initiates a verification call within 30 seconds
      </Step>

      <Step title="Code Entry">
        Enter the verification code provided during the automated call
      </Step>

      <Step title="Status Confirmation">
        Verify the caller ID status shows as "Verified" and ready for use
      </Step>
    </Steps>

    <Info>
      **Automatic Process**: Twilio automatically calls the provided number with a verification code - no manual intervention required.
    </Info>
  </Tab>

  <Tab title="Advanced Management">
    **Comprehensive Status Tracking:**

    * **✅ Verified**: Ready for outbound calls
    * **⏳ Pending**: Awaiting verification completion
    * **❌ Failed**: Verification failed, requires retry
    * **🔄 Processing**: Verification in progress

    **Management Actions:**

    * **Check Status**: Real-time validation with Twilio
    * **Update Status**: Refresh verification state
    * **Retry Verification**: Restart failed verification process
    * **Delete**: Remove caller ID from agent configuration
  </Tab>

  <Tab title="Verification Troubleshooting">
    <Warning>
      **Code Timing**: Verification codes are time-limited. Complete the process promptly after receiving the automated call.
    </Warning>

    **Common Solutions:**

    * Ensure the phone number can receive automated calls
    * Check for carrier restrictions on robotic calls
    * Verify number format is correct (+1XXXXXXXXXX for US)
    * Use "Retry Verification" for failed attempts
    * Allow up to 2 minutes for verification call delivery
  </Tab>
</Tabs>

## Agent Skills and Professional Expertise

Transform your agent with specialized industry skills and expertise templates that enhance conversation quality and task completion:

<Tabs>
  <Tab title="Skills Library">
    **Professional Specializations:**

    * **Customer Service Excellence**: Advanced support and resolution techniques
    * **Sales & Lead Qualification**: Persuasion and conversion strategies
    * **Healthcare Coordination**: Medical appointment scheduling and patient communication
    * **Real Estate Assistance**: Property information and showing coordination
    * **Financial Services Support**: Account management and financial guidance
    * **Legal Intake Specialization**: Case evaluation and consultation scheduling
    * **Technical Support Expert**: Troubleshooting and technical guidance

    **Enhanced Communication Skills:**

    * **Multi-language Proficiency**: Advanced cross-cultural communication
    * **Conflict Resolution**: De-escalation and problem-solving techniques
    * **Appointment Management**: Scheduling optimization and coordination
    * **Data Collection & Validation**: Information gathering and verification
  </Tab>

  <Tab title="Skill Selection Process">
    <Steps>
      <Step title="Browse Specialized Templates">
        Explore industry-specific skill templates with audio previews
      </Step>

      <Step title="Listen to Demonstrations">
        Play audio samples to understand each skill's communication style
      </Step>

      <Step title="Select Relevant Skills">
        Choose 3-5 skills that align with your agent's primary objectives
      </Step>

      <Step title="Automatic Integration">
        Skills integrate seamlessly with foundational components and conversation patterns
      </Step>
    </Steps>

    <Info>
      **Audio Previews**: Each skill template includes professional audio demonstrations showing real conversation examples.
    </Info>
  </Tab>

  <Tab title="Skill Optimization">
    <Info>
      **Focused Selection**: Choose skills that directly support your agent's core use case for maximum effectiveness.
    </Info>

    <Info>
      **Synergistic Combinations**: Select complementary skills that work together to create a comprehensive agent expertise profile.
    </Info>

    <Warning>
      **Avoid Skill Overload**: Too many diverse skills can dilute performance. Focus on your primary business objectives.
    </Warning>
  </Tab>
</Tabs>

## Voice Agents and Personality Selection

Choose from specialized voice agents with unique personalities and professional backgrounds designed for specific business scenarios:

<Tabs>
  <Tab title="Voice Agent Library">
    **Available Specializations:**

    * **Customer Service Representatives**: Professional, helpful, and solution-oriented
    * **Sales Professionals**: Persuasive, confident, and relationship-building focused
    * **Technical Support Specialists**: Patient, knowledgeable, and detail-oriented
    * **Healthcare Coordinators**: Empathetic, professional, and scheduling-focused
    * **Appointment Schedulers**: Efficient, organized, and calendar-management expert
    * **General Purpose Assistants**: Versatile, adaptable, and multi-skilled

    **Personality Characteristics:**

    * Unique speaking patterns and conversation styles
    * Industry-appropriate vocabulary and terminology
    * Professional demeanor and communication approaches
    * Specialized knowledge and expertise areas
  </Tab>

  <Tab title="Agent Selection Process">
    <Steps>
      <Step title="Browse Agent Profiles">
        Explore voice agents filtered by specialty and characteristics
      </Step>

      <Step title="Audio Preview">
        Listen to conversation samples to understand each agent's personality
      </Step>

      <Step title="Filter by Specialty">
        Use specialty filters to find agents matching your use case
      </Step>

      <Step title="Select and Configure">
        Choose an agent that aligns with your brand voice and objectives
      </Step>
    </Steps>

    <Info>
      **Interactive Previews**: All voice agents include high-quality audio samples demonstrating their unique conversation styles and personalities.
    </Info>
  </Tab>

  <Tab title="Agent Customization">
    **Voice Agent Benefits:**

    * Pre-configured personality traits and speaking patterns
    * Industry-optimized conversation flows and responses
    * Professional backgrounds and expertise integration
    * Consistent character maintenance across conversations

    <Info>
      **Seamless Integration**: Voice agents automatically configure appropriate voice settings, speech patterns, and conversational behaviors.
    </Info>
  </Tab>
</Tabs>

## Language and Voice Configuration

### Advanced Language Selection

Configure your agent's primary communication language with comprehensive regional support:

**Supported Languages:**

* 🇺🇸 **English** - Global business standard
* 🇪🇸 **Spanish** - Iberian and Latin American variants
* 🇫🇷 **French** - Metropolitan and Canadian French
* 🇩🇪 **German** - Standard and Austrian variants
* 🇮🇹 **Italian** - Standard Italian
* 🇵🇹 **Portuguese** - European Portuguese
* 🇧🇷 **Portuguese (Brazil)** - Brazilian variant
* 🇨🇳 **Chinese** - Mandarin (Simplified)
* 🇮🇳 **Hindi** - Standard Hindi
* 🇯🇵 **Japanese** - Standard Japanese
* 🇰🇷 **Korean** - Standard Korean
* 🇦🇷 **Arabic** - Modern Standard Arabic
* And **20+ additional languages** including Ukrainian, Vietnamese, Indonesian, and more

**Language Impact Areas:**

* Voice synthesis pronunciation and accent accuracy
* Speech recognition precision and processing
* Greeting template availability and suggestions
* AI model language comprehension and response quality
* Cultural communication patterns and etiquette

### Professional Voice Selection

<Tabs>
  <Tab title="Voice Library">
    **Comprehensive Voice Collection:**

    * **All Voices**: Complete professional voice library
    * **Male Voices**: Authoritative and confident options
    * **Female Voices**: Professional and approachable selections
    * **Search Function**: Find specific voices by name or characteristics

    **Voice Preview Features:**

    * High-quality audio samples for every voice
    * Context-appropriate conversation examples
    * Professional phone conversation demonstrations
    * Quality and clarity assessment tools
  </Tab>

  <Tab title="Voice Filtering">
    **Advanced Filter Options:**

    | Filter         | Purpose                      | Recommended Use             |
    | -------------- | ---------------------------- | --------------------------- |
    | **All Voices** | Browse complete library      | Initial exploration         |
    | **Male**       | Male voice characteristics   | Specific brand requirements |
    | **Female**     | Female voice characteristics | Audience preferences        |
    | **Search**     | Find specific voices         | Targeted voice selection    |

    <Info>
      **Quality Assurance**: All voices are professionally tested for phone conversation clarity and business communication appropriateness.
    </Info>
  </Tab>

  <Tab title="Voice Optimization">
    **Fine-tuning Parameters:**

    | Parameter            | Range         | Optimal Setting | Purpose                             |
    | -------------------- | ------------- | --------------- | ----------------------------------- |
    | **Speed**            | 0.70x - 1.20x | **1.06x**       | Natural conversation pace           |
    | **Stability**        | 0.0 - 1.0     | **0.15**        | Voice consistency across utterances |
    | **Similarity Boost** | 0.0 - 1.0     | **0.65**        | Balance naturalness and clarity     |

    <Info>
      **Production-Optimized Defaults**: Default settings (Speed: 1.06x, Stability: 0.15, Similarity Boost: 0.65) are optimized based on thousands of successful phone interactions.
    </Info>

    <Warning>
      **Parameter Adjustment**: Modifying voice settings requires thorough testing to ensure optimal conversation quality and user experience.
    </Warning>
  </Tab>
</Tabs>

## Advanced AI Model Selection

Choose from leading AI models across multiple providers to power your agent's intelligence and conversation capabilities:

<Tabs>
  <Tab title="Google Gemini (Recommended)">
    | Model                             | Performance | Best For                                | Status         |
    | --------------------------------- | ----------- | --------------------------------------- | -------------- |
    | **Gemini 3.1 Flash Lite Preview** | Optimized   | Balanced performance and reliability    | **🌟 Default** |
    | **Gemini 2.5 Flash**              | Advanced    | Complex reasoning and problem-solving   | Premium        |
    | **Gemini 2.5 Flash Lite**         | Efficient   | Cost-effective with strong capabilities | Balanced       |
    | **Gemini 2.0 Flash**              | Fast        | Quick responses for simple interactions | Speed-focused  |
    | **Gemini 1.5 Pro**                | Premium     | Advanced business logic and reasoning   | High-end       |

    **Gemini Advantages:**

    * **Optimized Default**: Gemini 3.1 Flash Lite Preview provides excellent out-of-the-box performance
    * **Advanced Context**: Superior conversation memory and context understanding
    * **Cost Efficiency**: Multiple options balancing capability and operational costs
    * **Reliability**: Proven performance across diverse business scenarios
  </Tab>

  <Tab title="OpenAI GPT">
    | Model            | Performance | Characteristics                   | Best For                    |
    | ---------------- | ----------- | --------------------------------- | --------------------------- |
    | **GPT-4o**       | High        | Balanced intelligence and speed   | General business use        |
    | **GPT-4.1**      | Premium     | Latest capabilities and reasoning | Complex problem-solving     |
    | **GPT-4.1 Mini** | Balanced    | Optimized efficiency              | Cost-effective operations   |
    | **GPT-4.1 Nano** | Ultra-fast  | Minimal latency responses         | Speed-critical applications |

    **GPT Selection Guide:**

    * **GPT-4.1**: Choose for complex scenarios requiring advanced reasoning
    * **GPT-4o**: Reliable choice for most business conversational needs
    * **GPT-4.1 Mini**: Optimal balance of capability, cost, and performance
    * **GPT-4.1 Nano**: Select when response speed is the primary priority
  </Tab>

  <Tab title="Anthropic Claude">
    **Claude Haiku 4.5**

    * **Performance**: Ultra-fast response times with minimal latency
    * **Best For**: Cost-effective interactions and straightforward conversations
    * **Use Cases**: Basic customer service, simple FAQ responses, high-volume operations
    * **Advantages**: Lowest operational costs, excellent speed, reliable performance

    <Info>
      **Cost Leadership**: Claude Haiku 4.5 provides excellent conversational capability at the most economical pricing tier.
    </Info>
  </Tab>

  <Tab title="Model Selection Strategy">
    <Info>
      **Speed Priority**: For phone conversations where response time is critical:

      * GPT-4.1 Nano (ultra-fast responses)
      * Gemini 2.0 Flash (rapid processing)
      * Claude Haiku 4.5 (minimal latency)
    </Info>

    <Info>
      **Intelligence Priority**: For complex conversations requiring advanced reasoning:

      * GPT-4.1 (latest capabilities)
      * Gemini 1.5 Pro (premium reasoning)
      * Gemini 2.5 Flash (advanced logic)
    </Info>

    <Info>
      **Balanced Performance**: For optimal capability-to-cost-to-speed ratio:

      * **Gemini 3.1 Flash Lite Preview** (recommended default)
      * GPT-4.1 Mini (efficient premium)
      * Gemini 2.5 Flash Lite (balanced performance)
    </Info>

    <Warning>
      **Model Impact**: Changing AI models significantly affects conversation patterns, response quality, and operational costs. Test thoroughly before deployment.
    </Warning>
  </Tab>
</Tabs>

## Meetzy Foundation Components

Leverage Meetzy's powerful built-in foundational components that provide professional-grade conversation capabilities refined through thousands of real-world interactions:

<Tabs>
  <Tab title="Enhanced Safety Guardrails">
    **Comprehensive Protection Framework:**

    * **Content Filtering**: Advanced inappropriate content prevention across all topics
    * **Professional Standards**: Consistent business communication tone and approach
    * **Edge Case Management**: Graceful handling of unexpected inputs and scenarios
    * **Security Protection**: Defense against prompt injection and manipulation attempts
    * **Compliance Assurance**: Adherence to industry communication standards
    * **Context Awareness**: Appropriate responses based on conversation context
    * **Industry Best Practices**: Following established business communication protocols

    **Configuration:**

    * **Default State**: Enabled by default for maximum protection
    * **Zero Setup**: Functions automatically without additional configuration
    * **Transparent Operation**: Works seamlessly with custom prompts and behaviors
    * **Adaptive Intelligence**: Automatically adjusts based on context and language

    <Warning>
      **Recommended Practice**: Keep safety guardrails enabled unless specific requirements conflict with professional conversation standards.
    </Warning>
  </Tab>

  <Tab title="Advanced Conversational Aspects">
    **Sophisticated Dialogue Management:**

    * **Natural Flow Control**: Advanced conversation pacing and rhythm optimization
    * **Active Listening Techniques**: Professional understanding and response patterns
    * **Intelligent Response Timing**: Context-aware pausing and conversation timing
    * **Empathetic Communication**: Emotional intelligence and appropriate emotional responses
    * **Professional Etiquette**: Industry-standard phone protocols and business courtesy
    * **Seamless Transitions**: Smooth topic changes and conversation redirection
    * **Cultural Sensitivity**: Appropriate communication styles for different regions
    * **Conflict Resolution**: Professional de-escalation and problem-solving approaches

    **Performance Benefits:**

    * **Immediate Professionalism**: Agents sound experienced from first interaction
    * **Accelerated Development**: Skip months of conversation optimization
    * **Consistent Excellence**: Every agent benefits from accumulated expertise
    * **Adaptive Intelligence**: Automatically adapts to specific business contexts

    <Info>
      **Battle-Tested Patterns**: Conversational aspects represent months of optimization based on successful real-world conversations.
    </Info>
  </Tab>

  <Tab title="Professional Agent Skills">
    **Comprehensive Professional Background:**

    * **Industry-Specific Knowledge**: Tailored professional contexts for various sectors
    * **Advanced Communication Skills**: Professional-level interaction capabilities
    * **Business Process Understanding**: Deep comprehension of operations and terminology
    * **Customer Interaction Mastery**: Extensive experience with diverse service scenarios
    * **Protocol Expertise**: Professional communication and business protocol understanding
    * **Situational Awareness**: Context-appropriate responses based on business needs
    * **Problem-Solving Experience**: Track record of resolving customer challenges
    * **Relationship Management**: Professional relationship building and maintenance skills

    **Specialized Skills Examples:**

    * **Sales Representatives**: Quota achievement and conversion expertise
    * **Customer Service Agents**: Resolution and satisfaction specialization
    * **Healthcare Coordinators**: Medical office and patient interaction experience
    * **Financial Services**: Compliance knowledge and account management
    * **Real Estate Professionals**: Market expertise and process knowledge
    * **Technical Support**: Troubleshooting and system knowledge specialization
  </Tab>

  <Tab title="Foundation Integration">
    **Synergistic Component Operation:**

    | Component                     | Primary Function          | Integration Level | Business Impact |
    | ----------------------------- | ------------------------- | ----------------- | --------------- |
    | **Safety Guardrails**         | Protective framework      | **Critical**      | Risk mitigation |
    | **Conversational Aspects**    | Natural communication     | **Essential**     | User experience |
    | **Professional Agent Skills** | Credibility establishment | **Important**     | Trust building  |

    <Info>
      **Synergistic Design**: All foundational components work together to create a cohesive, professional agent experience that adapts to your specific business requirements.
    </Info>

    <Info>
      **Accumulated Expertise**: These components represent industry best practices refined through analysis of thousands of successful conversations across multiple business sectors.
    </Info>

    <Warning>
      **Strong Recommendation**: Keep all foundational components enabled unless you have specific custom requirements that directly conflict. These are optimized for maximum performance.
    </Warning>
  </Tab>
</Tabs>

## Advanced Conversational Behavior Configuration

### Response Management

Fine-tune how your agent handles conversation timing and responses:

<Tabs>
  <Tab title="Response Settings">
    **Eagerness Control:**

    * **Low**: More patient, allows longer pauses before responding
    * **Normal**: Balanced response timing for natural conversation flow (recommended)
    * **High**: Quick responses, minimal wait time
    * **Auto**: AI-determined optimal timing based on context

    **Timeout Configuration:**

    * **Turn Timeout**: Time to wait before considering caller finished speaking (6 seconds default)
    * **Silence End Call Timeout**: Maximum silence before ending call (25 seconds default)
    * **Spelling Patience**: Handle spelling requests (auto/enabled/disabled)

    <Info>
      **Default Optimization**: Standard settings (Normal eagerness, 6s turn timeout, 25s silence timeout) are optimized for professional phone conversations.
    </Info>
  </Tab>

  <Tab title="Call Duration Management">
    **Duration Controls:**

    * **Maximum Duration**: Automatic call termination limit (600 seconds default)
    * **End Call Message**: Custom message when duration limit is reached
    * **Speculative Turn**: Enable predictive response preparation
    * **Background Speech Filter**: Filter out background noise and conversations

    **Advanced Features:**

    * Real-time conversation monitoring
    * Automatic quality adjustments
    * Dynamic timeout adaptation
    * Context-aware behavior modification

    <Warning>
      **Duration Limits**: Set appropriate maximum call durations to manage costs while allowing sufficient time for complex conversations.
    </Warning>
  </Tab>
</Tabs>

## Call Recording and Compliance

### Recording Configuration

**Comprehensive Call Recording Benefits:**

* **Quality Assurance**: Monitor and maintain high service standards
* **Training Data**: Collect valuable interaction data for improvement
* **Compliance Documentation**: Maintain audit trails and regulatory compliance
* **Performance Analytics**: Analyze conversation patterns and success rates
* **Customer Service Verification**: Validate service quality and resolution effectiveness

<Info>
  **Default Enabled**: Call recording is enabled by default to provide maximum value from every interaction and support continuous improvement.
</Info>

### Agent Documentation

**Internal Description Management:**

* **Maximum Length**: Up to 1,000 characters for comprehensive documentation
* **Internal Use**: Description is not visible to callers or external users
* **Team Coordination**: Help team members understand agent purpose and configuration
* **Organization**: Useful for managing multiple agents across different use cases
* **Deployment Notes**: Include special considerations and configuration details

## Integration and Advanced Editor Workflow

### Recommended Configuration Sequence

<Steps>
  <Step title="Foundation Setup">
    Complete all basic settings including name, company, phone configuration, and foundational components
  </Step>

  <Step title="Capability Selection">
    Choose specialized skills, voice agents, and AI models appropriate for your use case
  </Step>

  <Step title="Voice Optimization">
    Configure language, voice selection, and fine-tune voice parameters for optimal performance
  </Step>

  <Step title="Behavioral Configuration">
    Set up conversational behavior, response timing, and call management parameters
  </Step>

  <Step title="Verification and Testing">
    Verify all settings, test caller ID functionality, and validate configuration completeness
  </Step>

  <Step title="Advanced Configuration">
    Proceed to configure greetings, system prompts, and additional tools and actions
  </Step>
</Steps>

### Integration with Advanced Editor

**Seamless Workflow Integration:**

* **Voice Agents**: Automatically configure appropriate voice and behavioral settings
* **Specialized Skills**: Integrate with system prompts and conversation patterns
* **Foundation Components**: Work harmoniously with custom prompts and tools
* **Language Selection**: Influence greeting templates and conversation suggestions
* **AI Models**: Optimize for specific conversation complexity and requirements

<Info>
  **Workflow Optimization**: Complete basic settings configuration before proceeding to other Advanced Editor sections for optimal agent development experience.
</Info>

## Troubleshooting and Best Practices

<Tabs>
  <Tab title="Phone Configuration Issues">
    **Phone Number Problems:**

    <Warning>
      **Permanent Assignment**: Phone numbers cannot be changed after agent creation. Plan allocation carefully and verify availability before proceeding.
    </Warning>

    **Common Solutions:**

    * Verify Twilio account permissions and active status
    * Confirm phone number availability in your target regions
    * Check account credits and billing status
    * Validate webhook configurations are properly established
    * Ensure selected account has appropriate regional coverage
  </Tab>

  <Tab title="Caller ID Management">
    **Verification Challenges:**

    * **Delayed Verification Calls**: Allow up to 2 minutes for Twilio to initiate calls
    * **Invalid Codes**: Use "Retry Verification" for failed attempts
    * **Number Format Issues**: Ensure correct international format (+1XXXXXXXXXX for US)
    * **Carrier Restrictions**: Check if carrier blocks automated verification calls
    * **International Numbers**: Some regions may have verification limitations

    **Best Practices:**

    * Verify phone can receive calls during verification process
    * Complete verification promptly after receiving code
    * Use "Check Status" to confirm verification state
    * Maintain updated caller ID list for outbound campaigns
  </Tab>

  <Tab title="Performance Optimization">
    **Model Selection Strategy:**

    <Info>
      **Testing Approach**: Use the Advanced Editor Playground to test different AI models with your specific conversation scenarios before deployment.
    </Info>

    **Voice Configuration:**

    <Info>
      **Parameter Testing**: Test voice parameter adjustments with real conversation samples to ensure optimal quality and naturalness.
    </Info>

    **Foundation Components:**

    <Info>
      **Component Benefits**: Keep Meetzy foundation components enabled to leverage professional-grade conversation patterns and safety measures.
    </Info>

    <Warning>
      **Change Impact**: Any modifications to AI models, voice settings, or foundational components should be thoroughly tested before production deployment.
    </Warning>
  </Tab>
</Tabs>

## Advanced Features and Optimization

### Multi-Agent Management

**Administrative Capabilities:**

* **Company Assignment**: Organize agents by business unit or client
* **Twilio Account Selection**: Manage multiple telephony configurations
* **Resource Allocation**: Distribute phone numbers and caller IDs efficiently
* **Permission Management**: Control access and modification rights
* **Bulk Configuration**: Streamline setup for multiple similar agents

### Performance Monitoring

**Configuration Quality Assurance:**

* **Validation Checks**: Automated verification of configuration completeness
* **Component Integration**: Ensure all selected features work harmoniously
* **Resource Verification**: Confirm all required resources are properly assigned
* **Testing Preparation**: Configuration readiness for Playground testing
* **Deployment Validation**: Pre-deployment configuration verification

## Next Steps in Agent Development

<Steps>
  <Step title="Complete Basic Configuration">
    Finalize all basic settings, verify configurations, and ensure all required resources are properly assigned
  </Step>

  <Step title="Configure Agent Greetings">
    Set up appropriate inbound and outbound greetings that align with your agent's personality and use case
  </Step>

  <Step title="Define System Prompt">
    Create comprehensive personality and behavioral instructions that leverage your selected foundational components
  </Step>

  <Step title="Set Up Tools and Actions">
    Configure call handling, transfer capabilities, and external system integrations as needed
  </Step>

  <Step title="Comprehensive Testing">
    Use the Playground to thoroughly test all configurations and conversation scenarios
  </Step>

  <Step title="Deploy and Monitor">
    Launch your agent and continuously monitor performance using analytics and call recordings
  </Step>
</Steps>

<Info>
  **Continuous Improvement**: After deployment, regularly analyze call recordings and performance metrics to identify optimization opportunities and enhance agent effectiveness.
</Info>


# Copilot
Source: https://docs.meetzy.es/advanced-editor/copilot

AI-powered system engineer that transforms your agent with intelligent prompt editing, parameter configuration, rich visual diagnostics, and comprehensive setup

## What is Copilot?

Copilot is an advanced AI-powered system engineer integrated into the Advanced Editor. It goes far beyond simple prompt editing - **Copilot can configure your entire agent system** including input parameters, output settings, evaluations, audio expressiveness, and core configurations. Using sophisticated structured editing and intelligent analysis, Copilot transforms your agent based on natural language instructions and real call data.

<Info>
  Copilot has evolved from a prompt assistant to a complete system engineer - like having Data from Star Trek, but specifically designed for agent configuration. It can modify prompts, configure parameters, set up evaluations, update basic settings, adjust audio expressiveness tags, create and update tests from call history, and even suggest improvements - all through natural conversation and targeted operations. Beyond plain text, Copilot can also reply with **rich visual blocks** - stat panels, tables, charts, and cards - to make diagnostics and analysis instantly scannable.
</Info>

## Accessing Copilot

Copilot is located in the right panel of the Advanced Editor. Click the "Copilot" tab or press **⌘J** to access it quickly.

## Meet the Copilot Orb

Copilot now has a face: a gradient orb avatar that represents Meetzy across the entire chat experience, replacing the generic bot icon.

<Tabs>
  <Tab title="Idle State">
    A **slow, ambient breathing animation** while Copilot is waiting for your next message - a calm, always-present visual anchor in the empty state and next to every assistant reply.
  </Tab>

  <Tab title="Active State">
    When Copilot is thinking or streaming a response, the orb switches to a **faster pulse with expanding rings**, echoing the feeling of a call connecting - a subtle nod to Meetzy being a voice-agent copilot rather than a generic chat bot.
  </Tab>
</Tabs>

<Info>
  The orb appears in three places: the welcome screen hero, the avatar next to each assistant message, and the typing/thinking indicator - giving Copilot a consistent, recognizable identity throughout the conversation. Its state (**idle** or **active**) always reflects whether Copilot is waiting for you or actively working, so you always know at a glance whether it's inactive or engaged.
</Info>

## Enhanced Welcome Experience

When you first open Copilot, you'll see a clean welcome screen led by the Copilot orb, followed by **four quick action cards** to get started immediately without typing a single word:

<Tabs>
  <Tab title="Quick Action Cards">
    **Four color-coded cards** for the most common engineering workflows:

    * 🟣 **Diagnose performance** — analyzes real metrics, failed evaluations, and relevant calls, then shows the key findings visually (KPIs & call evidence)
    * 🌹 **Find recurring failures** — scans failed evaluations across your call history and returns a table of recurring issues with real examples
    * 🟢 **Create a regression test** — turns a real failed call into an automated regression test so the same bug never ships again
    * 🟡 **Improve and verify** — updates the prompt based on the current diagnosis and suggests how to verify the change afterward
  </Tab>

  <Tab title="One-Click Setup">
    **Instant analysis and fixes** with a single click:

    ```
    Click "Diagnose performance" → Runs a full performance diagnosis with visual KPIs
    Click "Find recurring failures" → Surfaces failure patterns in a table with call evidence
    Click "Create a regression test" → Builds a regression test from a real failed call
    Click "Improve and verify" → Improves the prompt and proposes a verification plan
    ```

    Each card pre-fills an optimized prompt and starts the conversation immediately - no typing required to get your first insight.
  </Tab>
</Tabs>

<Info>
  These quick actions are designed to close the loop: diagnose → find failures → fix with a regression test → improve and verify. Together they form a fast, guided workflow for continuously improving your agent.
</Info>

## Response Output Types

Copilot's replies are no longer limited to plain text. Depending on what you ask, Copilot decides the most effective format for its answer and can mix multiple output types in a single response:

<Tabs>
  <Tab title="Plain Text">
    **Conversational explanations** streamed in real time - used for reasoning, recommendations, step-by-step instructions, and any answer that reads best as natural language.
  </Tab>

  <Tab title="Rich Content Blocks">
    **Structured visual components** rendered inline with the chat - used whenever data is easier to scan than to read:

    * 📊 **Stat grids** for KPIs and metrics
    * 📋 **Tables** for lists of calls, failures, or test results
    * 📈 **Charts** for trends and comparisons
    * 🗂️ **Cards** for summaries, recommendations, and warnings

    Rich blocks always render alongside supporting text, so Copilot explains *what it found* in words and *shows the evidence* visually in the same message.
  </Tab>
</Tabs>

## Rich Response Blocks

Copilot doesn't just reply with text anymore - when the situation calls for it, it renders **rich, interactive blocks** directly inside the chat, making diagnostics and analysis far easier to scan than a wall of text. These blocks stream in the same way as regular messages, so you see them appear live as Copilot builds its answer.

<Tabs>
  <Tab title="Stat Grids">
    **At-a-glance KPI cards** for quick diagnostics:

    * Compact grid of metric cards (e.g. success rate, average duration, failed evaluations)
    * Optional **delta and trend indicators** (up, down, flat) to show change over time
    * Color-coded tones (neutral, info, success, warning, danger) to highlight what needs attention
    * Optional hint text under each stat for extra context

    Perfect for summarizing "Diagnose performance" results in a single scannable view.
  </Tab>

  <Tab title="Tables">
    **Structured data views** for lists of calls, failures, or test results:

    * Sticky header with clearly labeled columns
    * Automatic badge styling for common status values (Yes/No, Passed/Failed, Success/Error, Pending)
    * Scrollable, expandable body for long result sets without breaking the chat layout
    * Optional footnote for extra context or data source

    Ideal for "Find recurring failures" results, showing each failure pattern with examples side by side.
  </Tab>

  <Tab title="Charts">
    **Visual trend analysis** rendered inline with the conversation:

    * Supports line, bar, pie, and doughnut chart types
    * Multiple datasets with automatic color palette
    * Dark mode aware styling for axes, legends, and gridlines

    Great for visualizing performance trends over time or comparing metrics across segments.
  </Tab>

  <Tab title="Cards">
    **Highlighted callouts** for summaries, recommendations, or warnings:

    * Optional title, body text, and bullet list
    * Tone-based styling (neutral, info, success, warning, danger) to match the message
    * Used for wrapping up findings or proposing next steps after a diagnosis

    Often appears after a stat grid or table to summarize "what this means" and "what to do next."
  </Tab>
</Tabs>

<Info>
  Rich blocks stream in alongside regular text, so Copilot can explain a finding in natural language and immediately back it up with a stat grid, table, chart, or card - all within the same response. Look for these blocks whenever you ask Copilot to diagnose, compare, or summarize data - they turn a long paragraph into something you can scan in seconds.
</Info>

## System Engineering Capabilities

Copilot now operates as a comprehensive system engineer that can modify every aspect of your agent configuration:

<Tabs>
  <Tab title="Prompt Engineering">
    **Advanced Structured Editing**: Precise modifications using the EditPlan system

    * **replace\_section**: Completely replaces targeted sections with improved content
    * **insert\_after**: Adds new content after specified sections or markers
    * **append\_section**: Adds entirely new sections to the end of your prompt
    * **delete\_section**: Removes outdated or redundant sections cleanly

    Uses intelligent boundary detection to preserve document structure while making surgical improvements.
  </Tab>

  <Tab title="Parameter Configuration">
    **Intelligent Input & Output Setup** based on conversation analysis

    <Steps>
      <Step title="Input Parameters">
        Automatically suggests relevant input fields:

        * Role-based parameters (calendar preferences for schedulers)
        * Call-driven parameters (missing info from transcripts)
        * Industry-standard fields for your agent's domain
      </Step>

      <Step title="Output Configuration">
        Contextual output field suggestions:

        * Contact preferences (email, phone, text)
        * Lead qualification scores
        * Next step requirements
        * Custom fields based on conversation patterns
      </Step>

      <Step title="Output Enrichment">
        Comprehensive data processing instructions for captured information
      </Step>
    </Steps>
  </Tab>

  <Tab title="Evaluation & Testing">
    **Performance Measurement & Quality Assurance**

    * **Conversation Quality**: Clarity, professionalism, engagement metrics
    * **Goal Achievement**: Primary objective completion tracking
    * **Customer Satisfaction**: Experience and satisfaction evaluation
    * **Process Adherence**: Compliance with required procedures
    * **Test Generation**: Creates automated tests from call scenarios and history
    * **Test Updates**: Modifies existing tests in place - refining success conditions, chat history, or dynamic variables without creating duplicates
  </Tab>

  <Tab title="Basic Configuration">
    **Core Agent Settings** from natural language instructions

    * Agent name optimization based on role and personality
    * Inbound/outbound greeting configuration for different scenarios
    * Phone number assignment and management
    * Voice and personality adjustments for optimal customer experience
  </Tab>

  <Tab title="Audio Expressiveness">
    **Direct Control Over Voice Emotion Tags** (Eleven v3)

    * Adds, modifies, or removes expressive audio tags (e.g. `[excited]`, `[whispers]`, `[curious]`) directly in the prompt
    * Suggests contextually appropriate tags based on the agent's role and tone
    * Keeps tag usage consistent across the prompt to avoid overacting or flat delivery
    * Applies changes through the same purple-highlight review flow as other edits
  </Tab>
</Tabs>

## Enhanced SSE Protocol

Copilot uses a streamlined event system for real-time feedback across all system modifications:

| Event Type             | Description                                   | Visual Indicator                       | Applies To             |
| ---------------------- | --------------------------------------------- | -------------------------------------- | ---------------------- |
| **thinking**           | AI analyzing your request                     | ⏱️ Timer with seconds counter          | All operations         |
| **text**               | Streaming explanation content                 | 📝 Real-time text appearing            | All operations         |
| **edit**               | Structured prompt modification                | ✏️ Diff preview with operation details | Prompt changes         |
| **input\_change**      | Input parameter modification                  | 🔧 Parameter change card               | Input parameters       |
| **output\_change**     | Output setting update                         | 📤 Output configuration card           | Output settings        |
| **enrichment\_set**    | Output enrichment update                      | ⚡ Enrichment content preview           | Output enrichment      |
| **evaluation\_change** | Evaluation criteria update                    | 📊 Evaluation configuration            | Evaluations            |
| **test\_create**       | Automated test generation                     | 🧪 Test scenario preview               | Test creation          |
| **test\_update**       | Existing test modification                    | 🔁 Test diff preview                   | Test updates           |
| **agent\_config**      | Basic settings update                         | ⚙️ Configuration preview               | Core settings          |
| **audio\_tags\_set**   | Expressive audio tag update                   | 🎙️ Tag list preview                   | Audio expressiveness   |
| **rich\_block**        | Rich UI block (stat grid, table, chart, card) | 📊 Inline visual component             | Diagnostics & analysis |
| **complete**           | Operation finished                            | ✅ Summary with next steps              | All operations         |
| **error**              | Operation failed                              | ❌ Error with suggested fixes           | All operations         |

<Info>
  The `rich_block` event carries a `blockType` (`stat_grid`, `table`, `chart`, or `card`) plus a structured `data` payload. The frontend renders it as a native component inline with the rest of the streamed conversation, right next to any `text` events in the same response.
</Info>

## Visual Change Management System

### Purple Highlight Protocol

When Copilot generates configuration changes, distinctive visual indicators guide your review:

<Tabs>
  <Tab title="Pending Changes Banner">
    **Purple-highlighted banners** appear at the top of relevant panels:

    * **Copilot sparkle icon** with "Copilot Suggestions" title
    * **Change summary**: "3 new parameters, 1 evaluation"
    * **Bulk controls**: Accept All/Reject All buttons
    * **Panel navigation**: Direct links to affected configuration areas

    <Warning>
      Purple notification dots appear on affected panel tabs to ensure you don't miss suggestions.
    </Warning>
  </Tab>

  <Tab title="Change Cards">
    Each suggestion appears as a detailed **change card**:

    * **Operation badge**: Color-coded Add/Modify/Remove indicators
    * **Parameter details**: Name, type, description, and requirements
    * **Impact preview**: Shows how the change affects your agent
    * **Individual controls**: Accept/reject buttons for granular control
  </Tab>

  <Tab title="Cross-Panel Integration">
    **Seamless navigation** between Copilot and configuration panels:

    * Purple notification dots on affected panel tabs
    * Automatic panel highlighting when changes are generated
    * Smooth transitions preserve your editing context
    * Real-time sync between panels
  </Tab>
</Tabs>

### Change Type Indicators

| Change Type | Badge Color | Operation | Description                              |
| ----------- | ----------- | --------- | ---------------------------------------- |
| **Add**     | 🟢 Green    | ADD       | New parameters, evaluations, or settings |
| **Modify**  | 🟡 Yellow   | MODIFY    | Changes to existing configurations       |
| **Remove**  | 🔴 Red      | REMOVE    | Deletion of outdated or redundant items  |

## Comprehensive System Configuration

Copilot intelligently suggests changes across all agent configuration areas:

<Tabs>
  <Tab title="Input Parameter Operations">
    **Smart Parameter Suggestions** based on comprehensive analysis:

    <Steps>
      <Step title="Contextual Generation">
        Copilot analyzes your agent's role and referenced calls to suggest relevant input parameters:

        ```
        @call:abc123 The customer mentioned their budget repeatedly 
        but we didn't capture it. Add budget tracking parameters.
        ```

        Results in contextually relevant budget range and validation parameters.
      </Step>

      <Step title="Change Operations">
        Three types of parameter changes:

        * 🟢 **Add**: New parameters with proper validation and descriptions
        * 🟡 **Modify**: Updates to existing parameter configurations
        * 🔴 **Remove**: Clean up outdated or redundant parameters
      </Step>

      <Step title="Visual Review System">
        **Purple highlighting** in Input Parameters panel shows pending changes with individual accept/reject controls.
      </Step>
    </Steps>
  </Tab>

  <Tab title="Output & Enrichment Configuration">
    **Intelligent Output Field & Processing Setup**:

    ```
    The agent collected the customer's name and interest level 
    but we're missing their preferred contact method. Add comprehensive 
    output fields and enrichment instructions.
    ```

    Copilot suggests:

    * Contextually relevant output parameters
    * Data enrichment instructions for processing captured information
    * Webhook configuration recommendations
    * Output validation and formatting rules
  </Tab>

  <Tab title="Evaluation Criteria Setup">
    **Performance Measurement Configuration** based on agent goals:

    ```
    Create evaluations to measure conversation quality, goal achievement, 
    and customer satisfaction for this appointment scheduling agent.
    ```

    Automatically generates:

    * **Conversation Quality**: Clarity, professionalism, engagement scoring
    * **Goal Achievement**: Appointment booking success tracking
    * **Customer Satisfaction**: Experience evaluation criteria
    * **Process Adherence**: Compliance with scheduling procedures
  </Tab>

  <Tab title="Basic Settings & Configuration">
    **Core Agent Setup** from natural language:

    ```
    Update the agent name to "Sarah - Appointment Scheduler" and 
    make the greetings more friendly while keeping them professional.
    ```

    Copilot can modify:

    * Agent name and identity
    * Inbound/outbound greeting prompts
    * Phone number assignment
    * Voice and personality settings
  </Tab>

  <Tab title="Audio Expressiveness Tags">
    **Direct Tag Editing** for more natural, emotionally-aware voice delivery:

    ```
    Add expressive tags so the agent sounds more empathetic when 
    handling complaints, and more upbeat during the closing.
    ```

    Copilot can:

    * Insert new expressiveness tags at the right points in the prompt
    * Modify existing tags that feel overused, misplaced, or mismatched with tone
    * Remove tags that conflict with the agent's intended personality
    * Keep tag density balanced so delivery doesn't sound exaggerated
  </Tab>
</Tabs>

## Advanced Test Creation & Updates from Call History

One of Copilot's most powerful features is generating - and maintaining - comprehensive test suites based on real call data:

<Tabs>
  <Tab title="Test Generation from Calls">
    **Transform Call History into Automated Tests**:

    ```
    @call:abc123 This call shows great objection handling. 
    Create a test to ensure the agent maintains this quality.
    ```

    Copilot automatically extracts:

    * **Chat History**: Complete conversation flow from the call
    * **Success Conditions**: What made this call successful
    * **Success Examples**: Specific agent responses that worked well
    * **Failure Examples**: Alternative responses to avoid
    * **Dynamic Variables**: Contextual information that can be varied in tests
  </Tab>

  <Tab title="Updating Existing Tests">
    **Refine tests in place instead of duplicating them**:

    ```
    Update the "Objection handling - pricing" test to include the 
    new discount policy and tighten the success condition.
    ```

    Copilot can modify an existing test's:

    * **Name** and description
    * **Chat history** to reflect updated conversation flows
    * **Success condition** as your requirements evolve
    * **Success/failure examples** to match current expectations
    * **Dynamic variables** to cover new scenarios

    <Info>
      Test updates apply to the specific test referenced, so your test suite stays clean and versioned instead of accumulating near-duplicate tests over time.
    </Info>
  </Tab>

  <Tab title="Test Types Generated">
    **Multiple Test Scenarios** from single calls:

    * **Scenario Tests**: Complete conversation workflows
    * **Simulation Tests**: Realistic customer interactions
    * **Regression Tests**: Ensure quality doesn't degrade
    * **Edge Case Tests**: Handle unusual customer behaviors
  </Tab>

  <Tab title="Intelligent Test Structure">
    **Comprehensive Test Configuration**:

    * **Name**: Descriptive test identification
    * **Type**: Scenario or simulation classification
    * **Chat History**: Full conversation with timestamps
    * **Success Conditions**: Clear pass/fail criteria
    * **Dynamic Variables**: Customizable test parameters
    * **Quality Benchmarks**: Performance expectations
  </Tab>
</Tabs>

## Voice Input & Call Context Analysis

Enhanced natural interaction capabilities for comprehensive system configuration:

<Tabs>
  <Tab title="Voice Command System">
    <Steps>
      <Step title="Activation Methods">
        **Two ways to start voice recording**:

        * **Hold microphone button** in the input field
        * **Double-press Shift** to begin hands-free recording
      </Step>

      <Step title="Natural Speech Processing">
        **Speak conversationally about any system changes**:

        * "Add input parameters for customer budget and timeline preferences"
        * "Create evaluations to measure appointment booking success"
        * "Update the greetings to be more friendly and professional"
        * "Based on call A-B-C-1-2-3, improve the pricing objection handling"
      </Step>

      <Step title="System-Wide Understanding">
        **Copilot interprets requests across all configuration areas**:

        * Voice converts to text automatically
        * AI determines which systems need updates
        * Generates appropriate changes across multiple panels
      </Step>
    </Steps>
  </Tab>

  <Tab title="Call Context Integration">
    **Powerful call reference system** for data-driven improvements:

    ### Using Call References

    ```
    @call:abc123 The agent failed to capture the customer's budget 
    and preferred contact method. Add the missing parameters and 
    improve the data collection prompts.
    ```

    When you reference a call, Copilot:

    * Automatically loads the full transcript
    * Analyzes conversation flow and data gaps
    * Suggests missing input/output parameters
    * Updates prompts based on specific context
    * Creates relevant evaluation criteria
    * **Generates or updates tests based on call patterns**

    ### Comprehensive Call Analysis

    Referenced calls provide rich context for:

    * **Missing Data Identification**: What information wasn't captured
    * **Process Gap Analysis**: Where the conversation broke down
    * **Parameter Suggestions**: Input/output fields based on actual needs
    * **Evaluation Criteria**: Performance metrics based on real outcomes
    * **Test Creation & Updates**: New or refined automated test scenarios from actual interactions
  </Tab>
</Tabs>

## Accepting and Rejecting System Changes

### Granular Change Control Across All Panels

<Steps>
  <Step title="Cross-Panel Review">
    **Check all affected configuration areas**:

    * Look for purple notification dots on panel tabs
    * Review change cards in each affected area
    * Understand the impact of each suggested modification
  </Step>

  <Step title="Selective Application">
    **Accept only relevant changes across the system**:

    * Individual accept/reject buttons for each suggestion
    * Bulk accept/reject options within each panel
    * Changes apply immediately to respective configurations
  </Step>

  <Step title="Live System Updates">
    **See comprehensive results instantly**:

    * Input Parameters panel updates with new fields
    * Output Settings reflect new capture requirements
    * Evaluations panel shows new measurement criteria
    * Basic Settings display updated configurations
    * Tests panel includes newly generated or updated test scenarios
    * Prompt reflects updated audio expressiveness tags
  </Step>
</Steps>

### Comprehensive Change Management

**System-wide operations** for managing multiple configuration updates:

```
Accept All in Panel - Apply all suggestions within a specific configuration area
Accept All Changes - Apply all pending suggestions across the entire system
Reject All in Panel - Dismiss suggestions within a specific area
Reject All Changes - Dismiss all pending suggestions system-wide
```

<Warning>
  System-wide bulk actions affect multiple configuration areas simultaneously. Review individual suggestions across all panels before using "Accept All Changes".
</Warning>

## Advanced System Engineering Examples

### Comprehensive Agent Setup from Call Analysis

```
@call:xyz789 The agent struggled with budget questions, 
didn't capture contact preferences, and we can't measure success. 
Set up a complete configuration for this sales agent.
```

**Copilot generates**:

* **Prompt updates**: Better budget objection handling instructions
* **Input parameters**: Budget range, contact method preferences, lead source
* **Output parameters**: Qualification score, next steps, contact preferences
* **Evaluations**: Lead quality, objection handling, conversion tracking
* **Basic settings**: Professional but approachable greeting updates
* **Tests**: Automated scenarios based on the analyzed call patterns

### Multi-System Integration Configuration

```
Add comprehensive lead qualification for this real estate agent. 
Include budget tracking, timeline parameters, property preferences, 
qualification scoring, and success evaluations.
```

**Results in coordinated changes across**:

* Input Parameters: Budget, timeline, property type, location preferences
* Output Settings: Qualification score, lead temperature, next actions
* Prompt Instructions: Qualification process and scoring guidelines
* Evaluations: Lead quality metrics, conversion tracking, customer satisfaction
* Output Enrichment: Lead processing and CRM integration instructions
* Tests: Complete qualification workflow scenarios

### Performance Optimization from Multiple Calls

```
@call:call1 @call:call2 @call:call3 These calls show the agent 
needs better appointment scheduling process, missing confirmation 
parameters, and we can't track booking success rates.
```

**Comprehensive system updates**:

* Analyzes patterns across multiple call transcripts
* Updates prompts for consistent scheduling process
* Adds confirmation and calendar integration parameters
* Creates appointment booking success evaluations
* Suggests test scenarios based on real conversation patterns
* Generates regression tests for quality assurance

### Test Creation from Successful Calls

```
@call:perfect123 This was a perfect call where the agent handled 
all objections and booked the appointment. Create comprehensive 
tests to maintain this quality.
```

**Intelligent test generation**:

* Extracts complete conversation flow
* Identifies successful objection handling patterns
* Creates multiple test scenarios with variations
* Sets up quality benchmarks based on the successful call
* Includes dynamic variables for different customer personalities

### Updating a Test After a Policy Change

```
Update the "Refund policy objection" test - the refund window changed 
from 30 to 60 days, update the success condition and chat history accordingly.
```

**Copilot performs**:

* Locates the existing test by name
* Updates the success condition to reflect the new 60-day window
* Adjusts chat history examples that reference the old policy
* Leaves unrelated tests untouched

### Adjusting Audio Expressiveness for a More Natural Voice

```
The agent sounds too flat during the apology section. Add expressive 
tags there and remove the excessive excitement tags in the greeting.
```

**Copilot performs**:

* Inserts empathetic expressiveness tags (e.g. `[softly]`, `[apologetic]`) around the apology section
* Removes or tones down over-the-top tags in the greeting
* Shows the change through the purple-highlight diff review before applying

### Diagnosing Performance with Rich Blocks

```
Click "Diagnose performance" on the welcome screen
```

**Copilot responds with**:

* A short natural-language summary of the agent's current state
* A **stat grid** showing success rate, average call duration, and failed evaluations with trend indicators
* A **table** listing the most relevant calls tied to those metrics
* A closing **card** with tone-coded recommendations on what to fix first

### Finding Recurring Failures

```
Click "Find recurring failures" on the welcome screen
```

**Copilot responds with**:

* A **table** of recurring failure patterns extracted from failed evaluations
* Each row includes the failure type, frequency, and a linked call example
* A follow-up suggestion to turn the top failure into a regression test

### Visualizing Trends with a Chart

```
Show me how our success rate has changed over the last 4 weeks.
```

**Copilot responds with**:

* A short summary of the overall trend direction
* A **line chart** plotting success rate by week
* A closing note on whether the trend is improving, flat, or declining

## Quick Start with Predefined Actions

The welcome interface provides four **instant setup options**:

<Tabs>
  <Tab title="Diagnose Performance">
    **One-click performance diagnostics**:

    * Pulls real metrics, failed evaluations, and relevant calls
    * Renders findings as stat grids, tables, and charts
    * Closes with a card summarizing what needs attention

    Perfect for: Getting a fast health check on any agent
  </Tab>

  <Tab title="Find Recurring Failures">
    **Pattern detection across failed evaluations**:

    * Scans call history for repeated failure types
    * Presents results in a sortable, badge-styled table
    * Links directly to real call examples

    Perfect for: Root-causing quality issues before they scale
  </Tab>

  <Tab title="Create a Regression Test">
    **Turn bugs into safeguards**:

    * Builds a regression test from a real failed call
    * Captures the exact conditions that caused the failure
    * Adds it to your Tests panel to prevent recurrence

    Perfect for: Locking in a fix after resolving an issue
  </Tab>

  <Tab title="Improve and Verify">
    **Close the loop on prompt changes**:

    * Updates the prompt based on the latest diagnosis
    * Suggests a verification plan (Playground or regression test)
    * Keeps changes traceable through the diff review system

    Perfect for: Applying and validating a fix in one flow
  </Tab>
</Tabs>

## Next Step Suggestions & System Recommendations

After completing comprehensive changes, Copilot provides intelligent system-wide recommendations:

<Tabs>
  <Tab title="Related System Improvements">
    **Cross-system optimization suggestions**:

    * Added input parameters → suggests matching output fields and evaluations
    * Improved prompt handling → recommends relevant test creation or updates
    * Updated evaluations → suggests parameter adjustments for better measurement
    * Enhanced greetings → recommends matching closing procedures
    * Created or updated tests → suggests additional edge case scenarios
    * Adjusted audio tags → recommends verifying delivery in Playground
  </Tab>

  <Tab title="Configuration Completeness">
    **System audit recommendations**:

    * Missing evaluation criteria for new parameters
    * Suggested output enrichment for captured data
    * Recommended tests for new conversation flows
    * Configuration validation and optimization tips
    * Test coverage analysis and suggestions
  </Tab>

  <Tab title="Quality Assurance Prompts">
    **Comprehensive testing reminders**:

    * "Test parameter changes in Playground with real data"
    * "Create regression tests for updated conversation flows"
    * "Verify evaluation criteria with sample calls"
    * "Validate output enrichment with webhook testing"
    * "Run generated or updated tests to ensure quality maintenance"
    * "Listen to a Playground call to confirm the new audio tags sound natural"
  </Tab>
</Tabs>

## Conversation History & Iterative Configuration

Build sophisticated system improvements through multi-turn conversations:

```
User: Set up lead qualification for this sales agent

Copilot: [Creates input params, output fields, evaluations, and prompt updates]

User: Good, but make the budget qualification less aggressive

Copilot: [Refines parameters and prompt approach across affected areas]

User: Add follow-up scheduling and CRM integration

Copilot: [Adds scheduling params, webhook configs, and integration tests]

User: @call:abc123 Create tests based on this successful call

Copilot: [Generates comprehensive test scenarios from call history]

User: Update the "Follow-up scheduling" test - we now offer same-day slots

Copilot: [Updates the existing test's chat history and success condition in place]
```

<Info>
  Copilot maintains conversation context across multiple system modification requests, allowing for iterative refinement of your entire agent configuration through natural dialogue.
</Info>

## Comprehensive System Tasks

### Complete Agent Setup from Scratch

```
Create a professional appointment scheduling agent named "Sarah" 
with comprehensive lead qualification, appointment booking parameters, 
success tracking evaluations, friendly but professional greetings,
and test scenarios for quality assurance.
```

### Data-Driven System Optimization

```
@call:abc123 @call:def456 Analyze these problematic calls and 
optimize the entire agent system - fix the prompts, add missing 
parameters, create better evaluations, set up proper testing,
and generate quality assurance tests.
```

### Multi-System Integration Configuration

```
Set up this agent for CRM integration with Salesforce - add all 
necessary input and output parameters, webhook configurations, 
data enrichment instructions, integration success evaluations,
and comprehensive test scenarios for the integration workflow.
```

### Performance-Driven System Updates

```
Our conversion rate is low. Analyze the current setup and improve 
the qualification process, add better objection handling, create 
conversion tracking evaluations, optimize output data capture,
and generate tests to maintain improved performance.
```

### Test Suite Creation and Maintenance from Call History

```
@call:perfect1 @call:perfect2 @call:perfect3 These are our best calls.
Create a comprehensive test suite based on these successful patterns,
and update any existing tests that no longer match current policies.
```

### Compliance and Quality Assurance Setup

```
Add comprehensive compliance checking for this financial services 
agent - include required disclosures in prompts, compliance tracking 
parameters, regulatory evaluation criteria, audit trail outputs,
and compliance verification tests.
```

### Voice Expressiveness Tuning

```
Review the current audio tags for this customer support agent and 
adjust them so empathy comes through during complaints without 
sounding exaggerated anywhere else.
```

## Keyboard Shortcuts for System Navigation

| Shortcut          | Action                                |
| ----------------- | ------------------------------------- |
| **⌘J**            | Open/focus Copilot panel              |
| **Shift + Shift** | Start/stop voice recording            |
| **⌘Enter**        | Send message to Copilot               |
| **Escape**        | Close suggestions or cancel recording |
| **⌘1**            | Navigate to Basic Settings panel      |
| **⌘2**            | Navigate to Greetings panel           |
| **⌘3**            | Navigate to Input Parameters panel    |
| **⌘6**            | Navigate to Evaluations panel         |
| **⌘7**            | Navigate to Output Settings panel     |
| **⌘8**            | Navigate to Tests panel               |
| **⌘G**            | Navigate to Git Changes panel         |

## Best Practices for System Engineering

<Steps>
  <Step title="Start with Quick Actions">
    **Use the welcome-screen cards for common workflows**:

    * Start with "Diagnose performance" for a comprehensive health check
    * Use "Find recurring failures" to root-cause quality issues
    * Click "Create a regression test" to lock in a fix
    * Try "Improve and verify" to close the loop on a diagnosis
  </Step>

  <Step title="Use Comprehensive Natural Language">
    **Describe complete system needs rather than isolated changes**:

    **Good:** "Set up complete lead qualification with budget tracking, contact preferences, scoring system, success evaluations, and quality assurance tests"

    **Limited:** "Add a budget parameter"

    Comprehensive requests generate coordinated improvements across the entire system.
  </Step>

  <Step title="Leverage Real Call Data for System Design">
    **Reference actual calls to drive system-wide improvements**:

    * Use `@call:ID` syntax to provide specific examples of system gaps
    * Let Copilot analyze real conversation data across all configuration areas
    * Address actual problems with coordinated parameter, prompt, and evaluation updates
    * **Generate tests from successful calls, and update existing ones as requirements change**
  </Step>

  <Step title="Read the Rich Blocks Before Acting">
    **Use stat grids, tables, and charts to validate assumptions**:

    * Check trend indicators in stat grids before deciding what to prioritize
    * Scan failure tables for patterns rather than reacting to a single call
    * Use the closing card's recommendations as your next prompt to Copilot
    * Ask Copilot directly for a chart or table if a text answer feels hard to scan
  </Step>

  <Step title="Review All Affected Configuration Areas">
    **Check every panel with purple notification indicators**:

    * Input Parameters for new data collection fields
    * Output Settings for enhanced data capture
    * Evaluations for performance measurement
    * Basic Settings for core configuration updates
    * **Tests panel for newly generated or updated test scenarios**
    * Prompt for updated audio expressiveness tags
    * Consider the impact across the entire agent system
  </Step>

  <Step title="Think in System Integration Terms">
    **Consider how changes work together across the platform**:

    * New input parameters should have matching output fields
    * Prompt changes need corresponding evaluation criteria
    * Output configurations should include proper enrichment instructions
    * **Successful configurations should include test scenarios for quality assurance**
    * Test comprehensive workflows after system-wide changes
  </Step>

  <Step title="Use Voice for Complex System Requests">
    **Natural speech excels for describing comprehensive system needs**:

    * "Set up complete sales qualification with lead scoring, CRM integration, and quality tests"
    * "Based on these three problem calls, optimize the entire agent system with comprehensive testing"
    * Let conversational input guide holistic system improvements
  </Step>

  <Step title="Keep Tests Current Instead of Duplicating Them">
    **Update existing tests as your agent evolves**:

    * Reference an existing test by name to update its conditions instead of creating a near-duplicate
    * Use test updates after prompt or policy changes to keep your suite accurate
    * Reserve new test creation for genuinely new scenarios or call patterns
  </Step>

  <Step title="Fine-Tune Audio Expressiveness Deliberately">
    **Treat audio tags as part of the prompt, not an afterthought**:

    * Ask Copilot to review tag usage periodically, not just when something sounds wrong
    * Balance expressiveness so it enhances key moments without becoming distracting
    * Verify changes by listening to a Playground call after tags are updated
  </Step>

  <Step title="Follow Cross-Panel Notification Cues">
    **Purple indicators guide you to all affected system areas**:

    * Check every panel with notification dots
    * Don't miss configuration changes in different areas
    * **Review generated or updated tests in the Tests panel**
    * Use seamless navigation to review comprehensively
  </Step>

  <Step title="Test System Changes Comprehensively">
    **Validate complete system modifications before deploying**:

    * Test new parameters with real data in Playground
    * Verify evaluation criteria work with actual conversations
    * Validate output configurations and webhook integrations
    * **Run generated or updated tests to ensure quality maintenance**
    * Run regression tests across the entire conversation flow
  </Step>

  <Step title="Iterate on Complete System Understanding">
    **Build comprehensive agent configurations through dialogue**:

    * Start with high-level system requirements
    * Refine specific areas through follow-up conversations
    * Test and adjust system-wide configurations
    * **Generate and refine tests based on real performance**
    * Build understanding of how all components work together
  </Step>
</Steps>

## Next Steps

<CardGroup>
  <Card title="Input Parameters" icon="adjustments-horizontal" href="/advanced-editor/panels/input-parameters">
    Configure comprehensive input parameters with AI suggestions from Copilot
  </Card>

  <Card title="Output Settings" icon="arrow-up-tray" href="/advanced-editor/panels/output-settings">
    Set up output fields, enrichment, and webhooks with intelligent recommendations
  </Card>

  <Card title="Evaluations" icon="chart-bar" href="/advanced-editor/panels/evaluations">
    Create comprehensive performance measurements with Copilot's system analysis
  </Card>

  <Card title="Basic Settings" icon="cog-6-tooth" href="/advanced-editor/panels/basic-settings">
    Configure core agent settings with AI-powered optimization
  </Card>

  <Card title="Tests" icon="clipboard-check" href="/advanced-editor/tests">
    Create and maintain comprehensive automated tests for your entire agent system, including scenarios generated and updated from call history
  </Card>

  <Card title="Playground" icon="gamepad" href="/advanced-editor/playground">
    Test your complete system configuration with real data and validate generated test scenarios
  </Card>
</CardGroup>


# Evaluations
Source: https://docs.meetzy.es/advanced-editor/evaluations

Create automated quality tests for your AI agent

## What are Evaluations?

Evaluations are automated tests that assess your agent's performance after each call. They help you measure success rates, identify issues, and maintain quality standards across all conversations.

<Info>
  Evaluations provide quantitative metrics to track your agent's performance over time.
</Info>

## How Evaluations Work

Each evaluation is a **yes/no question** that the AI analyzes based on the conversation transcript. After every call, Meetzy runs all configured evaluations and stores the results.

### Example Evaluations

| Evaluation            | Question                                                |
| --------------------- | ------------------------------------------------------- |
| Appointment Scheduled | Did the customer schedule an appointment?               |
| Customer Satisfied    | Did the customer express satisfaction with the service? |
| Issue Resolved        | Was the customer's issue successfully resolved?         |
| Information Collected | Did the agent collect all required information?         |
| Objection Handled     | Did the agent successfully address customer objections? |

## AI Copilot Suggestions

<Info>
  The AI Copilot can automatically suggest performance metrics and evaluation criteria based on your agent's purpose and conversation patterns.
</Info>

### Automatic Evaluation Recommendations

The AI Copilot analyzes your agent configuration and suggests relevant evaluations:

1. **Context-Aware Suggestions**: Based on your agent's role (sales, support, appointments)
2. **Conversation Analysis**: Recommendations from analyzing test conversations
3. **Industry Best Practices**: Standard metrics for your use case
4. **Performance Optimization**: Evaluations that identify improvement opportunities

### Using Copilot Suggestions

<Steps>
  <Step title="Open AI Copilot">
    Click the sparkles icon in the right panel to access AI Copilot.
  </Step>

  <Step title="Request Evaluation Suggestions">
    Type something like "Suggest evaluations for my agent" or use the quick actions.
  </Step>

  <Step title="Review Recommendations">
    The Copilot will analyze your agent and suggest specific evaluation criteria.
  </Step>

  <Step title="Accept or Customize">
    Accept suggestions directly or modify them to fit your specific needs.
  </Step>
</Steps>

### Copilot Suggestion Types

<Tabs>
  <Tab title="Goal-Based">
    Evaluations focused on primary objectives:

    * "Did the agent achieve the call objective?"
    * "Was the primary goal accomplished?"
    * "Did the conversation meet success criteria?"
  </Tab>

  <Tab title="Quality Metrics">
    Performance and professionalism evaluations:

    * "Did the agent maintain professional tone?"
    * "Was the customer treated courteously?"
    * "Did the agent follow the conversation script?"
  </Tab>

  <Tab title="Compliance Checks">
    Regulatory and policy adherence:

    * "Were required disclosures provided?"
    * "Did the agent follow compliance protocols?"
    * "Was consent properly obtained?"
  </Tab>

  <Tab title="Data Collection">
    Information gathering effectiveness:

    * "Was all required information collected?"
    * "Did the agent verify customer details?"
    * "Were preferences and requirements captured?"
  </Tab>
</Tabs>

### Applying Copilot Suggestions

When the AI Copilot suggests evaluations, they appear in a purple suggestion banner:

<Warning>
  Review Copilot suggestions carefully before accepting them. While AI recommendations are based on best practices, they should be customized for your specific use case.
</Warning>

<Steps>
  <Step title="Review Suggestions">
    Examine the suggested evaluation name and question for relevance.
  </Step>

  <Step title="Customize as Needed">
    Modify the evaluation question to match your specific requirements.
  </Step>

  <Step title="Accept or Reject">
    Use the "Accept All" button to add all suggestions, or reject to dismiss them.
  </Step>

  <Step title="Fine-tune">
    Edit accepted evaluations in the main interface to perfect them.
  </Step>
</Steps>

## Creating Evaluations

<Steps>
  <Step title="Open Evaluations Section">
    Navigate to the Evaluations section in the left panel (✅ icon).
  </Step>

  <Step title="Add Evaluation">
    Click "Add Evaluation" to create a new test.
  </Step>

  <Step title="Enter Name">
    Give your evaluation a descriptive name.
  </Step>

  <Step title="Write Question">
    Create a clear yes/no question that can be answered from the transcript.
  </Step>

  <Step title="Save">
    Save the evaluation to add it to your agent.
  </Step>
</Steps>

### Evaluation Properties

| Property | Description          | Example                                                      |
| -------- | -------------------- | ------------------------------------------------------------ |
| Name     | Short identifier     | "Appointment Scheduled"                                      |
| Question | Yes/no prompt for AI | "Did the customer schedule an appointment during this call?" |

## Writing Effective Questions

<Info>
  The clearer your question, the more accurate the evaluation results will be.
</Info>

### Good Questions

* ✅ "Did the customer agree to schedule a follow-up call?"
* ✅ "Did the agent verify the customer's contact information?"
* ✅ "Was the customer's primary concern addressed during the call?"
* ✅ "Did the agent mention the current promotion?"

### Poor Questions

* ❌ "Was the call good?" (too vague)
* ❌ "How satisfied was the customer?" (not yes/no)
* ❌ "Rate the agent's performance" (not answerable from transcript)

### Copilot-Enhanced Question Writing

The AI Copilot can help improve your evaluation questions:

1. **Clarity Enhancement**: Makes vague questions more specific
2. **Binary Conversion**: Converts complex questions into yes/no format
3. **Context Addition**: Adds relevant context for better evaluation accuracy
4. **Best Practice Alignment**: Ensures questions follow evaluation best practices

## Using Templates

Meetzy provides pre-built evaluation templates for common use cases:

<Tabs>
  <Tab title="Sales">
    * "Did the customer express interest in the product?"
    * "Was pricing discussed during the call?"
    * "Did the agent attempt to close the sale?"
    * "Were all customer objections addressed?"
  </Tab>

  <Tab title="Support">
    * "Was the customer's issue resolved?"
    * "Did the agent follow the troubleshooting protocol?"
    * "Was an escalation necessary?"
    * "Did the customer confirm their issue was fixed?"
  </Tab>

  <Tab title="Appointments">
    * "Was an appointment successfully scheduled?"
    * "Did the agent confirm the appointment details?"
    * "Were alternative times offered if the preferred slot was unavailable?"
    * "Did the customer receive confirmation of the appointment?"
  </Tab>
</Tabs>

### Applying Templates

<Steps>
  <Step title="Click Templates">
    Click the "Templates" button in the Evaluations section.
  </Step>

  <Step title="Browse Categories">
    Select a category relevant to your use case.
  </Step>

  <Step title="Select Evaluations">
    Choose which evaluations to add.
  </Step>

  <Step title="Customize">
    Modify the questions to match your specific needs.
  </Step>
</Steps>

## Managing Evaluations

### Editing Evaluations

1. Click the edit icon (✏️) next to any evaluation
2. Modify the name or question
3. Save changes

### Deleting Evaluations

1. Click the delete icon (🗑️) next to the evaluation
2. Confirm deletion

<Warning>
  Deleting an evaluation removes it from future calls. Historical evaluation data is preserved.
</Warning>

### Reordering Evaluations

Drag and drop evaluations to change their display order.

## Viewing Results

Evaluation results are available in multiple places:

| Location            | What You See                                |
| ------------------- | ------------------------------------------- |
| Call Details        | Individual evaluation results for each call |
| Analytics Dashboard | Aggregated success rates over time          |
| Webhook Output      | Raw evaluation data sent to your systems    |
| API Response        | Programmatic access to results              |

## Best Practices

### Start with Copilot Suggestions

<Info>
  Begin by asking the AI Copilot to suggest evaluations. This provides a solid foundation based on your agent's configuration and industry best practices.
</Info>

### Evaluation Quality Guidelines

1. **Be Specific**: The more specific your question, the more reliable the evaluation
2. **Avoid Overlap**: Don't create multiple evaluations that measure the same thing
3. **Start Simple**: Begin with 3-5 core evaluations that measure your most important KPIs
4. **Review Regularly**: Periodically review evaluation results to calibrate questions
5. **Use for Training**: Low-scoring evaluations highlight areas for agent improvement

### Copilot Integration Tips

* **Iterate with AI**: Use the Copilot to refine and improve existing evaluations
* **Context Matters**: Provide context about your business when asking for suggestions
* **Test Suggestions**: Use the Playground to validate Copilot-suggested evaluations
* **Combine Sources**: Mix Copilot suggestions with templates and custom evaluations

## Common Evaluation Patterns

### Goal Achievement

```
Name: Primary Goal Met
Question: Did the call achieve its primary objective (sale, appointment, resolution, etc.)?
```

### Compliance

```
Name: Disclosure Made
Question: Did the agent provide the required legal disclosure at the beginning of the call?
```

### Quality

```
Name: Professional Tone
Question: Did the agent maintain a professional and courteous tone throughout the call?
```

### Data Collection

```
Name: Contact Info Verified
Question: Did the agent verify or collect the customer's email and phone number?
```

## Next Steps

<CardGroup>
  <Card title="AI Copilot" icon="sparkles" href="/advanced-editor/ai-copilot">
    Get AI-powered evaluation suggestions
  </Card>

  <Card title="Playground" icon="gamepad" href="/advanced-editor/playground">
    Test evaluations in real-time
  </Card>

  <Card title="Output Settings" icon="arrow-right-from-bracket" href="/advanced-editor/output-settings">
    Include evaluations in webhooks
  </Card>

  <Card title="Analytics" icon="chart-bar" href="/analytics">
    View evaluation results and trends
  </Card>
</CardGroup>


# Greetings
Source: https://docs.meetzy.es/advanced-editor/greetings

Configure welcome messages for inbound and outbound calls with multilingual templates

## What are Greetings?

Greetings are the initial messages your agent speaks when a call begins. Different greeting types are used depending on whether the call is inbound (customer calling you) or outbound (your agent calling the customer).

<Info>
  A well-crafted greeting sets the tone for the entire conversation and helps establish trust with callers.
</Info>

## Types of Greetings

### Inbound Greeting

Used when a customer calls your phone number. The agent should:

* Identify the company
* Welcome the caller
* Offer assistance

**Example:**

```
Hello, thank you for calling Acme Corp. My name is Sarah, 
how can I help you today?
```

### Outbound Greeting

Used when your agent initiates a call to a customer. The agent should:

* Introduce themselves
* State the purpose of the call
* Verify they're speaking with the right person

**Example:**

```
Hello, am I speaking with {{customer_name}}? 
Hi, this is Sarah from Acme Corp. I'm calling about 
your recent inquiry regarding our services.
```

## Multilingual Template System

The Advanced Editor provides pre-built greeting templates for multiple languages with country flags and categorized options. This system allows you to quickly apply professional greetings that match your agent's configured language.

### Available Languages

<Tabs>
  <Tab title="🇪🇸 Spanish">
    **Inbound Templates:**

    * "Gracias por llamar a \[Empresa]. ¿En qué puedo ayudarle?"
    * "Hola, ha contactado con \[Empresa]. ¿Cómo puedo asistirle?"
    * "Buenos días/tardes, habla con \[Empresa]. ¿En qué le puedo ayudar?"

    **Outbound Templates:**

    * "Hola, ¿hablo con \[Nombre]? Soy \[Agente] de \[Empresa]."
    * "Buenos días \[Nombre], le llamo de \[Empresa] en relación a..."
    * "Hola \[Nombre], soy \[Agente] y le contacto desde \[Empresa] porque..."
  </Tab>

  <Tab title="🇺🇸 English">
    **Inbound Templates:**

    * "Thank you for calling \[Company]. How may I help you today?"
    * "Hello, you've reached \[Company]. How can I assist you?"
    * "Good morning/afternoon, this is \[Company]. How may I help you?"

    **Outbound Templates:**

    * "Hello, may I speak with \[Name]? This is \[Agent] from \[Company]."
    * "Hi \[Name], this is \[Agent] calling from \[Company] regarding..."
    * "Good morning \[Name], I'm \[Agent] from \[Company] and I'm calling about..."
  </Tab>

  <Tab title="🇫🇷 French">
    **Inbound Templates:**

    * "Merci d'appeler \[Entreprise]. Comment puis-je vous aider?"
    * "Bonjour, vous avez joint \[Entreprise]. Comment puis-je vous assister?"
    * "Bonjour, vous êtes en ligne avec \[Entreprise]. Que puis-je faire pour vous?"

    **Outbound Templates:**

    * "Bonjour, puis-je parler à \[Nom]? Je suis \[Agent] de \[Entreprise]."
    * "Bonjour \[Nom], je vous appelle de \[Entreprise] concernant..."
    * "Bonjour \[Nom], c'est \[Agent] de \[Entreprise], je vous contacte au sujet de..."
  </Tab>

  <Tab title="🇩🇪 German">
    **Inbound Templates:**

    * "Vielen Dank für Ihren Anruf bei \[Unternehmen]. Wie kann ich Ihnen helfen?"
    * "Hallo, Sie haben \[Unternehmen] erreicht. Wie kann ich Sie unterstützen?"
    * "Guten Tag, hier ist \[Unternehmen]. Was kann ich für Sie tun?"

    **Outbound Templates:**

    * "Hallo, spreche ich mit \[Name]? Hier ist \[Agent] von \[Unternehmen]."
    * "Guten Tag \[Name], ich rufe von \[Unternehmen] an bezüglich..."
    * "Hallo \[Name], hier spricht \[Agent] von \[Unternehmen]. Ich rufe an wegen..."
  </Tab>

  <Tab title="🇮🇹 Italian">
    **Inbound Templates:**

    * "Grazie per aver chiamato \[Azienda]. Come posso aiutarla?"
    * "Salve, ha raggiunto \[Azienda]. Come posso assisterla?"
    * "Buongiorno, è in linea con \[Azienda]. In cosa posso aiutarla?"

    **Outbound Templates:**

    * "Salve, parlo con \[Nome]? Sono \[Agente] di \[Azienda]."
    * "Buongiorno \[Nome], la chiamo da \[Azienda] riguardo a..."
    * "Salve \[Nome], sono \[Agente] di \[Azienda] e la contatto per..."
  </Tab>

  <Tab title="🇵🇹 Portuguese">
    **Inbound Templates:**

    * "Obrigado por ligar para \[Empresa]. Como posso ajudá-lo?"
    * "Olá, você contatou \[Empresa]. Como posso assisti-lo?"
    * "Bom dia/tarde, você está falando com \[Empresa]. Em que posso ajudar?"

    **Outbound Templates:**

    * "Olá, falo com \[Nome]? Sou \[Agente] da \[Empresa]."
    * "Bom dia \[Nome], estou ligando da \[Empresa] sobre..."
    * "Olá \[Nome], sou \[Agente] da \[Empresa] e estou entrando em contato porque..."
  </Tab>
</Tabs>

## Using the Template System

<Steps>
  <Step title="Access Templates">
    In the Greetings section, click the "Templates" button next to either the Inbound or Outbound greeting field.
  </Step>

  <Step title="Browse Languages">
    A dropdown menu will appear showing all available languages with their respective country flags for easy identification.
  </Step>

  <Step title="Select Template">
    Choose the template that matches your agent's language configuration and call type (inbound/outbound).
  </Step>

  <Step title="Customize Parameters">
    The template will be inserted into the text field. Replace placeholder brackets like \[Company] and \[Agent] with your actual values or dynamic parameters.
  </Step>
</Steps>

<Warning>
  Ensure your chosen template matches your agent's configured language for optimal performance and natural conversation flow.
</Warning>

## Enhanced Dynamic Parameters

Make greetings personal and dynamic by including input parameters with enhanced functionality:

### Standard Parameters

```
Hello {{customer_name}}, this is Sarah from Acme Corp. 
I'm calling about your order {{order_number}} which is 
scheduled for delivery on {{delivery_date}}.
```

### Advanced Parameter Features

* **Automatic fallback**: Parameters gracefully handle missing values
* **Type-aware formatting**: Dates, numbers, and text are formatted appropriately
* **Conditional rendering**: Content adapts based on parameter availability

<Tip>
  Use the fallback phrase feature to ensure your greeting always works, even when dynamic parameters are unavailable.
</Tip>

## Fallback Phrases (Optional)

The new fallback system provides a safety net for your greetings:

### How Fallbacks Work

* **Primary greeting** uses dynamic parameters and full personalization
* **Fallback greeting** activates when any parameter is missing or empty
* **Seamless transition** ensures calls always start professionally

### Setting Up Fallbacks

<Steps>
  <Step title="Enable Fallback">
    Click the "Fallback phrase (optional)" toggle below your main greeting.
  </Step>

  <Step title="Write Simple Version">
    Create a simplified version of your greeting without dynamic parameters.
  </Step>

  <Step title="Test Scenarios">
    Your fallback will automatically activate when parameters are unavailable.
  </Step>
</Steps>

**Example Configuration:**

```
Primary: "Hello {{customer_name}}, this is {{agent_name}} from {{company}}..."
Fallback: "Hello! Thank you for calling. How can I help you today?"
```

## Configuring Greetings

<Steps>
  <Step title="Open Greetings Section">
    Navigate to the Greetings section in the left panel (👋 icon).
  </Step>

  <Step title="Choose Template or Custom">
    Either select a template from the dropdown or write a custom greeting.
  </Step>

  <Step title="Add Parameters">
    Include dynamic parameters using the  syntax.
  </Step>

  <Step title="Set Fallback (Optional)">
    Configure a fallback phrase for when parameters are unavailable.
  </Step>

  <Step title="Save Changes">
    Click Save to apply your greetings configuration.
  </Step>
</Steps>

## Language-Specific Best Practices

### Spanish Greetings

* Use formal "usted" form for professional calls
* Include time-appropriate greetings (buenos días/tardes)
* Maintain warm but respectful tone

### English Greetings

* Keep it concise and professional
* Use clear, simple language
* Include company identification early

### French Greetings

* Use appropriate formality level (vous/tu)
* Include proper salutations
* Maintain politeness throughout

### German Greetings

* Use formal address (Sie)
* Be direct but courteous
* Include proper titles when available

### Italian Greetings

* Use formal Lei form for business
* Include warm but professional tone
* Consider regional variations

### Portuguese Greetings

* Distinguish between PT and BR variants
* Use appropriate formality level
* Include welcoming language

## Greeting Tips

<AccordionGroup>
  <Accordion title="Match Agent Language" icon="language">
    Always ensure your greeting template matches your agent's configured language for natural conversation flow.
  </Accordion>

  <Accordion title="Keep It Brief" icon="hourglass-half">
    Long greetings can frustrate callers. Aim for 2-3 sentences maximum across all languages.
  </Accordion>

  <Accordion title="Use Fallbacks Wisely" icon="shield-check">
    Set up fallback phrases for critical dynamic parameters to ensure calls always start smoothly.
  </Accordion>

  <Accordion title="Test Parameters" icon="flask">
    Verify that all dynamic parameters work correctly before deploying your agent.
  </Accordion>

  <Accordion title="Cultural Sensitivity" icon="globe">
    Consider cultural norms and expectations for each language when crafting greetings.
  </Accordion>
</AccordionGroup>

## Best Practices by Call Type

### Inbound Calls

* ✅ Identify your company immediately
* ✅ Thank the caller for calling
* ✅ Use welcoming, helpful tone
* ✅ Offer assistance clearly
* ✅ Match the caller's language

### Outbound Calls

* ✅ Introduce yourself and company
* ✅ Verify you're speaking with the right person
* ✅ State the purpose briefly
* ✅ Be respectful of the recipient's time
* ✅ Use appropriate formality level

## Example Greetings by Industry

### Customer Support (Multilingual)

<Tabs>
  <Tab title="English">
    ```
    Thank you for calling {{company_name}} Support. 
    This is {{agent_name}}. How can I help you today?
    ```
  </Tab>

  <Tab title="Spanish">
    ```
    Gracias por llamar al soporte de {{company_name}}. 
    Habla {{agent_name}}. ¿En qué puedo ayudarle?
    ```
  </Tab>
</Tabs>

### Sales Calls (Multilingual)

<Tabs>
  <Tab title="English">
    ```
    Hello {{customer_name}}, this is {{agent_name}} from {{company_name}}. 
    I'm calling about your recent inquiry. Do you have a moment to chat?
    ```
  </Tab>

  <Tab title="French">
    ```
    Bonjour {{customer_name}}, c'est {{agent_name}} de {{company_name}}. 
    Je vous appelle concernant votre demande récente. Avez-vous un moment?
    ```
  </Tab>
</Tabs>

## Next Steps

<CardGroup>
  <Card title="System Prompt" icon="file-lines" href="/getting-started/create-agent#system-prompt">
    Define agent personality and behavior
  </Card>

  <Card title="Voice Configuration" icon="microphone" href="/advanced-editor/voice-settings">
    Configure voice and language settings
  </Card>

  <Card title="Dynamic Parameters" icon="code" href="/advanced-editor/parameters">
    Learn about parameter management
  </Card>

  <Card title="Evaluations" icon="check-double" href="/advanced-editor/evaluations">
    Test greeting effectiveness
  </Card>
</CardGroup>


# Input Parameters
Source: https://docs.meetzy.es/advanced-editor/input-params

Configure dynamic variables to personalize each call

## What are Input Parameters?

Input Parameters are dynamic variables that allow you to personalize each call with customer-specific information. When initiating an outbound call or receiving an inbound call through your API, you can pass these variables to customize the agent's behavior and knowledge.

<Info>
  Input Parameters are passed via API when initiating calls, making each conversation unique and personalized.
</Info>

## AI Copilot Suggestions

The AI Copilot can automatically analyze your agent's purpose and suggest relevant input parameters to enhance your call personalization. These suggestions appear as purple banners throughout the Input Parameters interface.

### How It Works

When you're working on your agent configuration, the AI Copilot:

* Analyzes your system prompt and agent purpose
* Identifies opportunities for personalization
* Suggests new parameters or modifications to existing ones
* Provides intelligent recommendations for parameter types and descriptions

### Suggestion Banners

AI suggestions appear in purple banners at the top of the Input Parameters section. Each banner contains:

* **Copilot icon** - Indicates this is an AI-generated suggestion
* **Summary** - Brief description of what changes are suggested
* **Parameter cards** - Detailed breakdown of each suggested parameter
* **Accept/Reject buttons** - Quick actions to apply or dismiss suggestions

### Managing Suggestions

<Tabs>
  <Tab title="Accept Suggestions">
    Click **Accept All** to apply all suggested parameters to your agent configuration. This will:

    * Add new suggested parameters
    * Modify existing parameters if improvements are suggested
    * Update parameter descriptions and types as recommended
  </Tab>

  <Tab title="Reject Suggestions">
    Click **Reject** to dismiss the current suggestions. You can:

    * Reject all suggestions at once
    * Continue working on your current configuration
    * Get new suggestions later as you modify your agent
  </Tab>
</Tabs>

<Warning>
  Accepting suggestions will modify your current input parameters. Make sure to review the changes before publishing your agent.
</Warning>

## Common Use Cases

| Use Case                | Example Parameters                                     |
| ----------------------- | ------------------------------------------------------ |
| Customer Identification | `customer_name`, `customer_id`                         |
| Appointment Booking     | `appointment_date`, `appointment_time`, `service_type` |
| Order Status            | `order_number`, `order_status`, `delivery_date`        |
| Account Management      | `account_balance`, `payment_due_date`                  |
| Lead Qualification      | `lead_source`, `product_interest`                      |

## Configuring Input Parameters

### Adding a Parameter

<Steps>
  <Step title="Open Input Params Section">
    Navigate to the Input Parameters section in the left panel.
  </Step>

  <Step title="Add New Parameter">
    Click "Add Parameter" to create a new variable.
  </Step>

  <Step title="Configure Properties">
    Set the name, type, and description.
  </Step>
</Steps>

### Parameter Properties

| Property      | Description                                 | Required |
| ------------- | ------------------------------------------- | -------- |
| Name          | Variable identifier (e.g., `customer_name`) | Yes      |
| Type          | Data type: string, number, boolean          | Yes      |
| Description   | What this parameter represents              | Yes      |
| Default Value | Fallback if not provided                    | No       |

### Data Types

<AccordionGroup>
  <Accordion title="String" icon="font">
    Text values like names, addresses, or identifiers.

    ```json theme={null}
    {
      "customer_name": "John Smith",
      "order_id": "ORD-12345"
    }
    ```
  </Accordion>

  <Accordion title="Number" icon="hashtag">
    Numeric values like amounts, quantities, or IDs.

    ```json theme={null}
    {
      "account_balance": 1250.50,
      "order_count": 3
    }
    ```
  </Accordion>

  <Accordion title="Boolean" icon="toggle-on">
    True/false values for flags and conditions.

    ```json theme={null}
    {
      "is_premium_customer": true,
      "has_pending_payment": false
    }
    ```
  </Accordion>
</AccordionGroup>

## Using Parameters in Prompts

Once defined, you can reference input parameters in your system prompt and greetings using the `{{parameter_name}}` syntax.

### Example: Personalized Greeting

**System Prompt:**

```
You are a customer service agent for Acme Corp.
You are speaking with {{customer_name}}.
Their account status is {{account_status}}.
```

**Greeting (Outbound):**

```
Hello {{customer_name}}, this is Sarah from Acme Corp. 
I'm calling about your order {{order_number}}.
```

### Example: Conditional Behavior

```
{{#if is_premium_customer}}
Treat this customer as a VIP. Offer priority support and exclusive deals.
{{else}}
Provide standard support. Mention upgrade options if relevant.
{{/if}}
```

## API Integration

When making API calls to initiate calls, include input parameters in your request:

```bash theme={null}
curl -X POST https://api.meetzy.io/v1/calls \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "agent_id": "agent_123",
    "phone_number": "+1234567890",
    "input_params": {
      "customer_name": "John Smith",
      "order_number": "ORD-12345",
      "is_premium_customer": true,
      "account_balance": 1250.50
    }
  }'
```

## Testing Input Parameters

Use the Playground to test how your agent handles different parameter values:

<Steps>
  <Step title="Open Playground">
    Navigate to the Playground in the right panel.
  </Step>

  <Step title="Open Context Variables">
    Click the "Context Variables" panel.
  </Step>

  <Step title="Set Test Values">
    Enter values for each input parameter.
  </Step>

  <Step title="Start Conversation">
    Test how the agent uses these values.
  </Step>
</Steps>

## Best Practices

<Tip>
  Use descriptive names and clear descriptions so the AI understands the context of each parameter.
</Tip>

1. **Use Clear Names** - `customer_name` is better than `cn`
2. **Provide Descriptions** - Help the AI understand context
3. **Set Default Values** - Prevent errors when parameters aren't provided
4. **Validate Before Sending** - Ensure parameters have valid values
5. **Document for Team** - Keep a reference of all parameters and their purposes
6. **Review AI Suggestions** - Let the Copilot help identify missing personalization opportunities

## Common Parameter Patterns

### Customer Information

```json theme={null}
{
  "customer_name": "string",
  "customer_phone": "string",
  "customer_email": "string",
  "customer_id": "string"
}
```

### Appointment Details

```json theme={null}
{
  "appointment_date": "string",
  "appointment_time": "string",
  "service_type": "string",
  "location": "string"
}
```

### Order Information

```json theme={null}
{
  "order_number": "string",
  "order_total": "number",
  "items_count": "number",
  "delivery_date": "string"
}
```

## Next Steps

<CardGroup>
  <Card title="Output Settings" icon="arrow-right-from-bracket" href="/advanced-editor/output-settings">
    Configure data extraction
  </Card>

  <Card title="Playground" icon="gamepad" href="/advanced-editor/playground">
    Test with different parameters
  </Card>
</CardGroup>


# Introduction to Advanced Editor
Source: https://docs.meetzy.es/advanced-editor/introduction

Learn about the Advanced Editor including Creation Overlay with prompt-first agent creation, Meetzy Foundation components, and specialized skills system

## What is the Advanced Editor?

The **Advanced Editor** is Meetzy's comprehensive agent configuration tool that gives you full control over every aspect of your AI agent's behavior. It now features a revolutionary **Creation Overlay** with prompt-first creation, Meetzy Foundation components, specialized skills, and enhanced team collaboration capabilities.

<Info>
  The Advanced Editor is the recommended tool for creating production-ready agents with custom behaviors, integrations, and advanced logic. The new Creation Overlay with prompt-first creation makes it faster than ever to get started.
</Info>

## Nueva Experiencia de Creación de Agentes

Una de las actualizaciones más significativas del Editor Avanzado es la **nueva experiencia de creación de agentes** que revoluciona cómo construyes asistentes de IA. Esta innovación introduce un enfoque completamente nuevo donde los usuarios pueden describir simplemente sus necesidades en lenguaje natural y obtener un agente pre-configurado como punto de partida.

### Cómo Funciona la Nueva Experiencia

<Steps>
  <Step title="Describe tu Agente en Lenguaje Natural">
    Simplemente describe qué quieres que haga tu agente en español o cualquier idioma

    Ejemplo: "Un agente que responde llamadas de atención al cliente para una clínica dental, agenda citas y transfiere casos urgentes a la recepcionista"
  </Step>

  <Step title="Generación Automática Inteligente">
    El sistema de IA de Meetzy analiza tu descripción y automáticamente:

    * **Configura ajustes básicos** del agente usando componentes Foundation
    * **Detecta e integra habilidades especializadas** según tus requisitos
    * **Genera mensajes de saludo** personalizados para tu caso de uso
    * **Optimiza prompts del sistema** específicos para tu dominio
    * **Selecciona voz y idioma** apropiados para tu región
    * **Prepara estructura de base de conocimiento** lista para tu contenido
  </Step>

  <Step title="Configuración Basada en Foundation">
    Construido sobre **Meetzy Foundation**, tu agente incluye:

    * Patrones de conversación específicos de la industria
    * Conocimiento de dominio pre-entrenado
    * Plantillas de respuesta optimizadas
    * Integraciones avanzadas de habilidades
  </Step>

  <Step title="Transición sin Interrupciones">
    La experiencia de creación transiciona suavemente al Editor Avanzado completo para personalización y ajuste fino de la configuración generada por Foundation
  </Step>
</Steps>

### Beneficios de la Nueva Experiencia

**Velocidad de Implementación**

* **Reducción de tiempo** del 80% en la configuración inicial
* **Eliminación de configuración técnica compleja** para empezar
* **Punto de partida inteligente** con mejores prácticas incorporadas
* **Enfoque en personalización** en lugar de configuración desde cero

**Precisión y Calidad**

* **Componentes Foundation probados** en miles de implementaciones
* **Detección automática de habilidades** basada en análisis de IA
* **Configuración optimizada** para casos de uso específicos
* **Patrones de conversación** validados por la industria

**Accesibilidad Mejorada**

* **Interfaz en lenguaje natural** accesible para usuarios no técnicos
* **Configuración automática inteligente** reduce curva de aprendizaje
* **Sugerencias contextuales** para mejorar la descripción
* **Soporte multiidioma** para descripciones en cualquier idioma

<Warning>
  Puedes saltar la nueva experiencia de creación y comenzar con un agente en blanco haciendo clic en "empezar desde cero" durante el proceso de creación.
</Warning>

## Creation Overlay - Revolutionary Prompt-First Agent Building

The Advanced Editor now begins with an elegant **Creation Overlay** that transforms how you build AI agents. This innovative feature introduces a **prompt-first approach** where users can simply describe what their agent should do, and Meetzy automatically builds it using Foundation components and specialized skills.

### How Creation Overlay Works

<Steps>
  <Step title="Describe Your Agent">
    Simply describe what you want your agent to do in plain English

    Example: "An agent that answers customer support calls for a dental clinic, books appointments, and transfers urgent cases to the receptionist"
  </Step>

  <Step title="AI-Powered Generation with Foundation Components">
    Meetzy's AI analyzes your description and automatically configures:

    * **Basic agent settings** and metadata using Foundation components
    * **Specialized skills** based on your agent's requirements
    * **Initial greeting messages** tailored to your use case
    * **System prompts** optimized for your specific domain
    * **Voice and language settings** appropriate for your region
    * **Knowledge base structure** ready for your content
  </Step>

  <Step title="Foundation-Powered Configuration">
    Built on **Meetzy Foundation**, your agent includes:

    * Industry-specific conversation patterns
    * Pre-trained domain knowledge
    * Optimized response templates
    * Advanced skill integrations
  </Step>

  <Step title="Seamless Transition">
    The overlay smoothly transitions to the full Advanced Editor for customization and fine-tuning of the generated Foundation setup
  </Step>
</Steps>

### Creation Overlay Features

**Beautiful Visual Experience**

* **WebGL wave animation** provides an immersive, professional background
* **Clean, distraction-free interface** focused entirely on creation
* **Responsive design** works perfectly on desktop and mobile devices

**Natural Language Processing with Foundation Integration**

* **Intelligent prompt analysis** understands complex agent requirements
* **Foundation component selection** automatically chooses appropriate base components
* **Skill identification** detects required specialized skills from your description
* **Context-aware configuration** sets appropriate defaults based on use case
* **Multi-language support** for describing agents in various languages

**Smart Defaults & Foundation Configuration**

* **Automatic voice selection** based on agent purpose and language
* **Foundation-powered greetings** using industry-specific templates
* **System prompt optimization** leveraging Foundation knowledge
* **Skill integration** automatic configuration of detected capabilities
* **Knowledge base preparation** with Foundation-structured templates

<Warning>
  You can always skip the Creation Overlay and start with a blank agent by clicking "or start from scratch" during the creation process.
</Warning>

### Meetzy Foundation Integration

The Creation Overlay leverages **Meetzy Foundation** - a comprehensive set of pre-built components that serve as the building blocks for AI agents:

**Foundation Components Include:**

* **Industry Templates**: Pre-configured setups for healthcare, real estate, retail, and more
* **Conversation Patterns**: Proven dialogue flows optimized for different business scenarios
* **Domain Knowledge**: Built-in understanding of industry-specific terminology and processes
* **Skill Libraries**: Specialized capabilities like appointment booking, information retrieval, and call routing

**Skills System Integration:**

* **Appointment Management**: Booking, rescheduling, and cancellation workflows
* **Customer Support**: Ticket creation, escalation, and knowledge base integration
* **Sales Assistance**: Lead qualification, product information, and CRM integration
* **Information Gathering**: Structured data collection and validation
* **Call Routing**: Intelligent transfer and escalation logic

### Creation Overlay Interface Elements

The Creation Overlay features a minimal, focused design:

* **Navigation**: Easy back button to return to the agents dashboard
* **Meetzy Branding**: Clean logo placement that adapts to light/dark themes
* **Central Prompt Area**: Large textarea with gradient border effects for describing your agent
* **Smart Send Button**: Appears automatically when text is entered
* **Foundation Indicator**: Visual feedback showing Foundation components being applied
* **Keyboard Shortcuts**: Enter key to generate, ESC to close
* **Skip Option**: Direct link to traditional editor for advanced users

## Enhanced Navigation System

### Integrated Agent Selector

The Advanced Editor includes a powerful **agent selector** directly in the toolbar, enabling seamless navigation between agents without leaving the editor environment.

<Info>
  The agent selector displays the current agent name as a dropdown button with visual avatars, making it easy to identify and switch between different agents quickly.
</Info>

#### Agent Selector Capabilities

* **Visual Agent Identification**: Avatar-based indicators with agent name initials
* **Quick Switching**: Dropdown interface for rapid agent navigation
* **Search Functionality**: Built-in search to find agents by name
* **Company Filtering**: Admin users can filter agents by company
* **Context Preservation**: Unsaved changes protection during navigation
* **Real-time Updates**: Automatically reflects current agent state

### Navigation Best Practices

<Warning>
  If you have unsaved changes when switching agents, you'll be prompted to save or discard them before proceeding to prevent data loss.
</Warning>

## When to Use the Advanced Editor

| Scenario                                      | Recommended Tool                     |
| --------------------------------------------- | ------------------------------------ |
| Creating a new agent from natural description | ✅ Advanced Editor (Creation Overlay) |
| Building Foundation-powered agents            | ✅ Advanced Editor (Creation Overlay) |
| Configuring specialized skills                | ✅ Advanced Editor (Full Editor)      |
| Fine-tuning agent responses and behavior      | ✅ Advanced Editor                    |
| Adding knowledge base documents               | ✅ Advanced Editor                    |
| Configuring webhook integrations              | ✅ Advanced Editor                    |
| Testing agents in real-time                   | ✅ Advanced Editor                    |
| Quick agent duplication                       | Agents Dashboard                     |
| Bulk agent management                         | Agents Dashboard                     |

## Interface Overview

The Advanced Editor features a professional two-panel layout with enhanced collaboration tools, Foundation integration indicators, and specialized skills management.

### Left Panel - Configuration

The left panel contains all configuration sections accessible via an intuitive icon bar:

| Icon | Section         | Description                                                    | Shortcut |
| ---- | --------------- | -------------------------------------------------------------- | -------- |
| ⚙️   | Basic Settings  | Name, language, voice, LLM model, Foundation setup             | ⌘1       |
| 👋   | Greetings       | Foundation-powered welcome messages for inbound/outbound calls | ⌘2       |
| 🔧   | Input Params    | Dynamic variables for personalized calls                       | ⌘3       |
| 📚   | Knowledge Base  | Documents, FAQs, Foundation knowledge                          | ⌘4       |
| 🛠️  | Actions & Tools | Skills, integrations, transfers, webhooks                      | ⌘5       |
| 📊   | Evaluations     | Automated quality tests and success criteria                   | ⌘6       |
| 📤   | Output Settings | Data extraction and webhooks                                   | ⌘7       |

### Right Panel - Testing & Tools

| Tab             | Purpose                                                         | Shortcut |
| --------------- | --------------------------------------------------------------- | -------- |
| ✨ Meetzy        | AI assistant to help improve your prompts and Foundation setup  | ⌘J       |
| 🎮 Playground   | Test your agent with Foundation components and skills           | ⇧⌘P      |
| 📋 Tests        | Create and run automated regression tests                       | ⇧⌘T      |
| 🔄 Changes      | Track changes with Git-like version control                     | ⌘G       |
| 📞 Recent Calls | Browse the agent's most recent calls without leaving the editor | —        |

<Info>
  The **Recent Calls** tab shows the last 30 days of calls for the agent you're currently editing, including status, success outcome, duration, and a quick summary. Click any call to open its full recording and transcript in a new tab, or use "View all" to jump to the Calls page filtered for this agent.
</Info>

## Focus Mode - Distraction-Free Editing

Sometimes you just need to concentrate on the **System Prompt** without the visual noise of configuration panels and testing tools around you. **Focus Mode** does exactly that: it hides the left configuration panel and the right panel (Meetzy Copilot, Playground, Tests, Changes, Recent Calls), leaving only the Monaco editor centered on your screen.

<Info>
  Toggle Focus Mode instantly with the keyboard shortcut **⌘\\** on macOS or **Ctrl+\\** on Windows/Linux. Press it again to bring the side panels back.
</Info>

### How Focus Mode Works

<Steps>
  <Step title="Trigger Focus Mode">
    Press **Cmd+\\** (macOS) or **Ctrl+\\** (Windows/Linux) anywhere in the Advanced Editor.
  </Step>

  <Step title="Panels Disappear">
    The left configuration panel and the right panel (Copilot, Playground, Tests, Changes, Recent Calls) smoothly collapse out of view.
  </Step>

  <Step title="Editor Takes Center Stage">
    The Monaco editor expands to use the freed-up space, giving you a clean, full-width writing surface for your system prompt.
  </Step>

  <Step title="Exit Anytime">
    Press the same shortcut again to instantly restore the panels and return to the standard layout.
  </Step>
</Steps>

### When to Use Focus Mode

* **Writing or rewriting long system prompts** where every distraction counts
* **Reviewing prompt structure** using the outline without side panel clutter
* **Screen-sharing or recording demos** where a clean interface looks more professional
* **Working on smaller screens** where side panels compete for space with the editor

<Warning>
  Focus Mode only hides the panels visually, it does not disable your unsaved changes protection, Git status tracking, or auto-save behavior. All your work is preserved exactly as before.
</Warning>

#### Expand the Right Panel for a Better Reading Experience

The right panel now includes an **expand button** that widens the panel so you can read Meetzy Copilot's messages, Playground results, Tests, Changes history, and Recent Calls much more comfortably.

<Info>
  Look for the expand icon in the right panel's header, right next to the tab selector. Clicking it widens the panel to a more spacious layout, and clicking it again returns it to its original size.
</Info>

It's a small addition, but once you toggle it on, you'll immediately notice how much easier it is to follow long Copilot explanations, compare Playground conversation turns, review detailed diffs in the Changes tab, or scan through Recent Calls, no more squinting at cramped text in a narrow column.

<Steps>
  <Step title="Open any right panel tab">
    Select Meetzy, Playground, Tests, Changes, or Recent Calls from the right panel tabs.
  </Step>

  <Step title="Click the expand icon">
    Locate the expand button in the top-right area of the panel and click it to widen the panel.
  </Step>

  <Step title="Read comfortably">
    Enjoy a wider, more readable layout for long messages, test results, change diffs, or call summaries.
  </Step>

  <Step title="Collapse when done">
    Click the icon again to collapse the panel back to its default width and regain space for the editor.
  </Step>
</Steps>

<Warning>
  On smaller screens, expanding the right panel may temporarily reduce the visible space for the left configuration panel and the Monaco editor. Collapse it again once you've finished reading.
</Warning>

<Info>
  Focus Mode and the expandable right panel work great together: use Focus Mode when you need to write without distractions, and expand the right panel when you need to read long Copilot answers, Playground transcripts, Changes diffs, or Recent Calls details.
</Info>

## Advanced Version Control System

The Advanced Editor includes a sophisticated **Git-like version control system** that transforms team collaboration on AI agent development. This enterprise-grade system provides real-time collaboration with automatic conflict detection.

<Info>
  The version control system enables seamless team collaboration with automatic conflict detection, branching capabilities, and comprehensive change tracking.
</Info>

### Git Status Indicators

The Advanced Editor displays comprehensive version control status in the top toolbar:

| Status    | Badge Color | Description                                    |
| --------- | ----------- | ---------------------------------------------- |
| Clean     | 🟢 Green    | No local changes, up to date with remote       |
| Modified  | 🟡 Yellow   | Local changes pending save (shows count)       |
| Conflicts | 🔴 Red      | Merge conflicts require resolution             |
| Syncing   | 🔵 Blue     | Currently synchronizing with remote repository |
| Locked    | 🟠 Orange   | Another team member is actively editing        |

<Warning>
  When status shows "Conflicts" or "Locked", resolve these issues before attempting to publish changes to avoid data loss.
</Warning>

### Collaborative Features

The Git-like system supports sophisticated team workflows:

* **Real-time collaboration awareness**: See when teammates are editing simultaneously
* **Automatic conflict detection** and guided resolution
* **Branch-based workflows** for parallel development
* **Comprehensive audit trails** with detailed change history
* **Rollback capabilities** to safely revert changes
* **Team coordination tools** to prevent edit conflicts

## Enhanced Testing System

### Improved Test Management

The testing system has been significantly enhanced with powerful new capabilities:

* **Foundation component testing**: Validate Foundation-powered features
* **Skills testing**: Test individual specialized skills
* **Batch testing**: Run multiple test iterations with configurable repeat counts (1-10x)
* **Test filtering**: Filter tests by status (All, Pass, Fail) with visual counters
* **Test history**: View comprehensive execution history with detailed results
* **Input parameters**: Configure dynamic variables for realistic test scenarios
* **Simulation mode**: Interactive conversation simulation for testing

### Advanced Test Features

<Tabs>
  <Tab title="Foundation Testing">
    Test Foundation components and their integration:

    * Validate industry-specific conversation patterns
    * Test Foundation knowledge application
    * Verify skill integrations work correctly
    * Analyze Foundation component performance
  </Tab>

  <Tab title="Skills Testing">
    Test individual specialized skills:

    * Appointment management workflows
    * Information gathering capabilities
    * Call routing and transfer logic
    * Integration with external systems
  </Tab>

  <Tab title="Scenario Tests">
    Test specific conversation scenarios with defined success criteria:

    * Configure user personas and conversation goals
    * Set multiple success/failure examples using Foundation patterns
    * Use dynamic variables for realistic scenarios
    * Track test performance over time
  </Tab>

  <Tab title="Interactive Simulation">
    Real-time conversation simulation:

    * Live conversation testing with Foundation components
    * Skills-enabled conversation flows
    * Configurable turn limits for focused testing
    * Real-time streaming of AI responses
    * Context variable injection for dynamic scenarios
  </Tab>

  <Tab title="Batch Reliability Testing">
    Run tests multiple times for statistical confidence:

    * Configurable repeat counts (1-10x)
    * Batch progress tracking with visual indicators
    * Aggregate results analysis
    * Statistical reliability reporting
  </Tab>
</Tabs>

### Test Status Indicators

The enhanced testing interface provides clear visual feedback:

* **Pass/Fail counters**: See test results at a glance with color-coded badges
* **Foundation indicators**: Show which Foundation components are being tested
* **Skills markers**: Identify which specialized skills are active in tests
* **Running indicators**: Animated status indicators for active tests
* **History tracking**: Easy access to previous test runs and results
* **Input parameters**: Visual indicators for configured dynamic variables

## Meetzy Foundation & Skills System

### Foundation Components

**Meetzy Foundation** provides industry-specific building blocks that power your agents:

**Healthcare Foundation**

* Medical terminology understanding
* Appointment booking workflows
* HIPAA-compliant conversation patterns
* Patient information handling

**Real Estate Foundation**

* Property inquiry handling
* Showing scheduling
* Lead qualification processes
* Market information delivery

**Retail Foundation**

* Product information delivery
* Order management workflows
* Customer support patterns
* Inventory inquiries

**Financial Services Foundation**

* Account inquiry handling
* Transaction support
* Compliance-aware conversations
* Secure information exchange

### Specialized Skills

The skills system provides plug-and-play capabilities:

**Core Skills**

* **Appointment Booking**: Complete scheduling workflows with calendar integration
* **Information Retrieval**: Intelligent search and knowledge base querying
* **Call Transfer**: Smart routing based on conversation context
* **Data Collection**: Structured information gathering with validation

**Advanced Skills**

* **CRM Integration**: Real-time customer data access and updates
* **Payment Processing**: Secure transaction handling
* **Multi-language Support**: Dynamic language switching
* **Sentiment Analysis**: Real-time mood detection and response adaptation

<Info>
  Skills are automatically suggested and configured based on your Creation Overlay prompt, but can be manually added or removed in the Actions & Tools section.
</Info>

## Workflow Best Practices

<Steps>
  <Step title="Start with Creation Overlay">
    Use the Creation Overlay to describe your agent in natural language. The AI will select appropriate Foundation components and configure specialized skills automatically.
  </Step>

  <Step title="Review Foundation Setup">
    Examine the automatically configured Foundation components and skills. Verify they match your requirements and add or remove as needed.
  </Step>

  <Step title="Check Version Control Status">
    Always verify the Git status indicator before starting work to understand collaboration state and avoid conflicts.
  </Step>

  <Step title="Pull Latest Team Changes">
    Use "Check" and "Pull" buttons to synchronize with your team's latest changes before making modifications.
  </Step>

  <Step title="Fine-tune Foundation Configuration">
    Adjust Foundation-powered settings, greetings, and system prompts generated from your initial description. Toggle Focus Mode (⌘\ / Ctrl+\\) if you want to write without distractions.
  </Step>

  <Step title="Configure Skills">
    Review and configure the specialized skills that were automatically added based on your prompt.
  </Step>

  <Step title="Add Knowledge & Tools">
    Upload knowledge base documents and configure additional integrations as needed for your specific use case.
  </Step>

  <Step title="Run Foundation Testing">
    Use the enhanced testing system to validate Foundation components, skills, and overall behavior comprehensively. Expand the right panel if you need more room to review long test results.
  </Step>

  <Step title="Check Recent Calls">
    Once your agent is live, use the Recent Calls tab to quickly review real conversations without leaving the editor, spotting issues or opportunities to refine your prompt.
  </Step>

  <Step title="Review & Commit Changes">
    Use the Git changes panel to review modifications and commit with descriptive messages for proper version control.
  </Step>

  <Step title="Deploy to Production">
    Safely publish your Foundation-powered agent once all tests pass and collaboration status is clean.
  </Step>
</Steps>

## Key Features & Benefits

### Prompt-First Creation Revolution

The Creation Overlay with Foundation integration transforms agent development:

**Natural Language Interface**

* Describe your agent's purpose in everyday language
* AI-powered analysis with Foundation component selection
* Automatic skill detection and configuration
* Intelligent Foundation template matching

**Foundation-Powered Generation**

* Industry-specific templates automatically applied
* Pre-configured conversation patterns
* Built-in domain knowledge integration
* Optimized system prompts based on Foundation best practices

**Time & Efficiency Savings**

* Skip complex technical configuration
* Start with battle-tested Foundation components
* Leverage pre-built skills and integrations
* Focus immediately on customization rather than setup

### Enhanced Foundation Integration

**Smart Component Selection**

* Automatic Foundation component matching based on description
* Industry-specific configuration optimization
* Pre-built conversation flows and patterns
* Domain knowledge integration

**Skills Auto-Configuration**

* Intelligent skill detection from your prompt
* Automatic integration setup
* Pre-configured workflows and templates
* Ready-to-use specialized capabilities

### Enhanced Navigation

The integrated agent selector revolutionizes workflow efficiency:

* **Seamless switching** between agents without losing context
* **Search functionality** to quickly locate specific agents
* **Visual identification** with avatar-based agent indicators
* **Context preservation** with comprehensive unsaved changes protection

### Focus Mode for Deep Work

Focus Mode is designed for the moments when you need to think clearly about your agent's behavior:

* **One shortcut, instant results**: ⌘\ (macOS) or Ctrl+\ (Windows/Linux) toggles the layout immediately
* **Zero configuration lost**: hiding panels never affects your unsaved changes, Git status, or auto-save
* **Full-width editing**: the Monaco editor expands to use all the available space
* **Toggle anytime**: press the shortcut again to bring back the left and right panels instantly

<Info>
  Focus Mode is a purely visual toggle, think of it as closing the door to your office. Everything you were working on is still there when you open it again.
</Info>

### A More Comfortable Right Panel

The new expand toggle for the right panel might look like a minor visual tweak, but in daily use it turns out to matter more than expected:

* **Wider reading area** for long Copilot explanations and suggestions
* **Easier comparison** of Playground conversation turns side by side
* **Clearer Tests review** when reading detailed pass/fail explanations
* **More readable diffs** in the Changes tab when reviewing version history
* **More scannable calls** when browsing Recent Calls entries and summaries

<Info>
  Like Schrödinger's cat, this feature seems unremarkable until you actually open the box, once you try expanding the panel while reading a long Copilot response, you'll likely keep using it every day.
</Info>

### Recent Calls, Right Where You Need It

Reviewing real conversations used to mean leaving the editor entirely and navigating to a separate Calls page. Now, the **Recent Calls** tab brings the last 30 days of activity for the agent you're editing directly into the right panel:

* **At-a-glance status**: color-coded badges for finished, failed, busy, ringing, and other call outcomes
* **Success indicators**: quickly spot calls marked as successful based on your evaluation criteria
* **Relative and absolute timestamps**: see "2 hours ago" at a glance, with the exact date and time available on hover
* **One-click access**: open any call's full recording and transcript in a new tab without losing your place in the editor
* **View all**: jump straight to the Calls page pre-filtered for the current agent when you need the complete history

<Info>
  Recent Calls refreshes on demand with a single click, so you can always check the very latest interactions while you're mid-edit on your system prompt.
</Info>

### Advanced Collaboration

The Git-like version control system brings enterprise-grade collaboration:

**Team Coordination**

* Real-time awareness of teammate editing activities
* Automatic conflict detection prevents data conflicts
* Lock indicators show when files are being edited by others

**Change Management**

* Comprehensive audit trails for all modifications
* Ability to revert to previous versions safely
* Branch-based workflows for feature development

**Production Safety**

* Staged deployment process prevents accidental changes
* Version validation before publishing ensures quality
* Rollback capabilities for quick issue resolution

## Next Steps

<CardGroup>
  <Card title="Creation Overlay" icon="sparkles" href="/advanced-editor/creation-overlay">
    Master the prompt-first Creation Overlay with Foundation components and automatic skill configuration
  </Card>

  <Card title="Foundation Components" icon="building-library" href="/advanced-editor/foundation-components">
    Learn about Meetzy Foundation components and how they power your agents
  </Card>

  <Card title="Skills System" icon="puzzle-piece" href="/advanced-editor/skills-system">
    Understand specialized skills and how to configure them for your use cases
  </Card>

  <Card title="Agent Navigation" icon="arrows-right-left" href="/advanced-editor/agent-selector">
    Learn to use the integrated agent selector for efficient workflow management
  </Card>

  <Card title="Version Control" icon="code-branch" href="/advanced-editor/version-control">
    Understand Git-like collaboration, conflict resolution, and team coordination
  </Card>

  <Card title="Enhanced Testing" icon="clipboard-check" href="/advanced-editor/tests">
    Use Foundation testing, skills validation, and advanced test features
  </Card>

  <Card title="Basic Settings" icon="gear" href="/advanced-editor/basic-settings">
    Configure Foundation-powered settings, voice, LLM model, and core configuration
  </Card>

  <Card title="Knowledge Base" icon="book" href="/advanced-editor/knowledge-base">
    Add documents, FAQs, and integrate Foundation knowledge with your content
  </Card>
</CardGroup>


# Knowledge Base
Source: https://docs.meetzy.es/advanced-editor/knowledge-base

Add documents, FAQs, and reference materials for your agent

## What is the Knowledge Base?

The Knowledge Base is where you provide your agent with information it can reference during conversations. This includes product documentation, FAQs, company policies, pricing information, and any other content your agent needs to answer customer questions accurately.

<Info>
  A well-populated Knowledge Base significantly improves your agent's ability to provide accurate, helpful responses by giving it access to your specific business information.
</Info>

## Adding Documents

You can add knowledge to your agent in three ways:

<Tabs>
  <Tab title="URL Import">
    Import content directly from web pages.

    <Steps>
      <Step title="Click 'Add URL'">
        Open the Knowledge Base section and select the URL option.
      </Step>

      <Step title="Enter the URL">
        Paste the full URL of the webpage you want to import.
      </Step>

      <Step title="Configure Settings">
        Set the document name and choose the usage mode.
      </Step>

      <Step title="Confirm Import">
        The system will fetch and process the content automatically.
      </Step>
    </Steps>

    **Best for:** Product pages, help center articles, documentation sites, pricing pages.
  </Tab>

  <Tab title="Text Entry">
    Manually enter or paste text content.

    <Steps>
      <Step title="Click 'Add Text'">
        Select the text entry option.
      </Step>

      <Step title="Enter Content">
        Type or paste your content in the text area.
      </Step>

      <Step title="Configure Settings">
        Set the document name and choose the usage mode.
      </Step>

      <Step title="Save">
        The content is processed and added to the knowledge base.
      </Step>
    </Steps>

    **Best for:** Custom FAQs, internal guidelines, scripts, policies.
  </Tab>

  <Tab title="File Upload">
    Upload documents directly from your computer.

    <Steps>
      <Step title="Click 'Add File'">
        Select the file upload option.
      </Step>

      <Step title="Select File">
        Choose a file from your computer (PDF, TXT, DOC, DOCX).
      </Step>

      <Step title="Configure Settings">
        Set the document name and choose the usage mode.
      </Step>

      <Step title="Upload">
        The document is processed and indexed automatically.
      </Step>
    </Steps>

    **Supported formats:** PDF, TXT, DOC, DOCX, and more.
  </Tab>
</Tabs>

## Usage Modes

Each document in your Knowledge Base can be configured with one of two usage modes:

| Mode            | Icon | Description                                                              | When to Use                                              |
| --------------- | ---- | ------------------------------------------------------------------------ | -------------------------------------------------------- |
| **Auto Search** | 🔍   | Agent searches this document only when relevant to the conversation      | Large documents, extensive catalogs, reference materials |
| **In Prompt**   | 📋   | Content is always included in the agent's context for every conversation | Critical info, company details, core policies            |

### Understanding Usage Modes

<Info>
  **Auto Search** uses semantic search to find relevant portions of documents when needed, making conversations more efficient and cost-effective.
</Info>

<Warning>
  **In Prompt** mode includes the entire document content in every conversation, which can increase costs and slow responses for large documents. Use only for essential, concise information.
</Warning>

#### Auto Search Mode

* **How it works:** The agent automatically searches through the document when it detects the conversation topic is relevant
* **Performance:** Efficient for large documents - only retrieves relevant sections
* **Cost:** Lower cost as only relevant portions are processed
* **Best practices:** Use for comprehensive guides, product catalogs, detailed FAQs

#### In Prompt Mode

* **How it works:** The document content is included in the agent's context for every single conversation
* **Performance:** Immediate access to information but can slow down responses
* **Cost:** Higher cost as content is processed in every conversation
* **Best practices:** Use only for critical information under 1000 characters

## Managing Documents

### Document List View

The Knowledge Base displays all your documents in a organized list showing:

| Column      | Description                                |
| ----------- | ------------------------------------------ |
| **Name**    | Document title (editable)                  |
| **Type**    | Visual indicator: 🔗 URL, 📄 Text, 📁 File |
| **Mode**    | Usage mode: 🔍 Auto Search or 📋 In Prompt |
| **Preview** | First few lines of content                 |
| **Actions** | Edit (✏️) and Delete (🗑️) buttons         |

### Editing Existing Documents

You can modify existing documents to update content, names, or usage modes using the edit functionality.

<Steps>
  <Step title="Click Edit Button">
    Locate the edit icon (✏️) next to any document in your Knowledge Base list.
  </Step>

  <Step title="Modify Document Fields">
    The edit modal will open with current document details loaded. You can modify different fields based on the document type:

    * **All documents:** Name and usage mode
    * **URL documents:** Source URL (will refetch content)
    * **Text documents:** Full text content
    * **File documents:** Only name and usage mode (content cannot be edited)
  </Step>

  <Step title="Save Changes">
    Click "Save Changes" to update the document. Changes are processed immediately.
  </Step>
</Steps>

#### Document Type-Specific Editing

<Tabs>
  <Tab title="URL Documents">
    **Editable Fields:**

    * Document name
    * Source URL
    * Usage mode (Auto Search / In Prompt)

    **How it works:** When you update the URL, the system will automatically refetch the content from the new location and update the document.

    **Use cases:** Updating documentation URLs, fixing broken links, switching to newer versions of web pages.
  </Tab>

  <Tab title="Text Documents">
    **Editable Fields:**

    * Document name
    * Full text content
    * Usage mode (Auto Search / In Prompt)

    **How it works:** You can modify the text content directly in the editor. The system loads the complete document content for editing.

    <Warning>
      When you edit the text content of a document, the system creates a new document ID internally. This ensures proper content indexing but means conversation references to the old document ID will not work.
    </Warning>

    **Use cases:** Updating FAQs, modifying policies, correcting information, adding new details.
  </Tab>

  <Tab title="File Documents">
    **Editable Fields:**

    * Document name
    * Usage mode (Auto Search / In Prompt)

    **Limitations:** The actual file content cannot be edited through the interface.

    **How it works:** You can only change the display name and how the document is used by the agent.

    **To update content:** Delete the document and upload a new version of the file.
  </Tab>
</Tabs>

### Real-Time Content Loading

For optimal performance, the system handles content loading intelligently:

* **Document List:** Shows content previews for quick reference
* **Edit Modal:** Loads full document content automatically when editing text or URL documents
* **Loading States:** Clear indicators when content is being fetched
* **Error Handling:** Graceful fallback to preview content if full loading fails

<Info>
  When editing URL or text documents, the system automatically loads the complete content. This may take a few seconds for large documents.
</Info>

### Document Deletion

<Steps>
  <Step title="Click Delete Icon">
    Click the delete icon (🗑️) next to the document.
  </Step>

  <Step title="Confirm Deletion">
    Confirm the action in the popup dialog.
  </Step>

  <Step title="Immediate Removal">
    The document is immediately removed from your agent's knowledge base.
  </Step>
</Steps>

<Warning>
  Deleted documents cannot be recovered. Make sure to export or backup important content before deleting.
</Warning>

## Best Practices

### Content Organization

* **Keep Current:** Regularly update documents to reflect latest information using the edit functionality
* **Be Specific:** Include exact phrases and terms customers might use
* **Structure Well:** Use clear headings, bullet points, and organized sections
* **Avoid Duplication:** Don't repeat the same information across multiple documents

### Editing Workflow

* **Review Before Editing:** Check the current content and usage mode before making changes
* **Test After Updates:** Verify the agent can access and use the updated information correctly
* **Monitor Performance:** Large content changes may affect response times
* **Consider Document IDs:** Remember that editing text content creates new document IDs

### Usage Mode Strategy

* **Auto Search for:** Large documents, comprehensive guides, detailed product info
* **In Prompt for:** Company name, key contact info, essential policies (keep under 1000 chars)

### Performance Optimization

* **Test Regularly:** Verify the agent can find and use information correctly
* **Monitor Conversations:** Review how the agent uses knowledge base content
* **Optimize Content:** Remove outdated or redundant information
* **Update Incrementally:** Make small, targeted edits rather than large overhauls

<Info>
  Start with Auto Search mode for most documents. Only switch to In Prompt mode for absolutely critical information that must always be available.
</Info>

## Example Knowledge Base Structure

Here's a recommended structure for organizing your knowledge base:

```
📚 Knowledge Base
├── 📄 Company Overview (In Prompt)
│   └── Essential company info, mission, key contacts
├── 🔗 Product Catalog URL (Auto Search)  
│   └── Complete product listings and specifications
├── 📄 Pricing Information (Auto Search)
│   └── Current pricing tiers and packages
├── 📁 FAQ Document (Auto Search)
│   └── Common customer questions and answers
├── 📄 Refund Policy (Auto Search)
│   └── Terms and conditions for returns
├── 🔗 Support Articles URL (Auto Search)
│   └── Help center and troubleshooting guides
└── 📄 Contact Information (In Prompt)
    └── Phone numbers, hours, emergency contacts
```

## Integration with Agent Context

Your Knowledge Base documents are dynamically integrated into your agent's conversational context:

* **Semantic Search:** Auto Search documents are searched based on conversation relevance
* **Context Injection:** In Prompt documents are always available in the agent's working memory
* **Real-Time Updates:** Changes to documents are immediately available to the agent
* **Intelligent Retrieval:** The system automatically selects the most relevant information

## Troubleshooting

### Common Issues

**Agent can't find information:**

* Verify the document is saved and shows in the Knowledge Base list
* Try using Auto Search mode for better semantic matching
* Include variations of keywords customers might use
* Check if recent edits created new document IDs

**Slow agent responses:**

* Check if you have too many In Prompt documents
* Consider moving large documents to Auto Search mode
* Review total character count of In Prompt content

**Outdated information:**

* Use the edit functionality to update documents directly
* Changes take effect immediately in new conversations
* Consider setting up regular review schedules

**Content loading issues:**

* Verify internet connectivity for URL documents
* Check file format compatibility for uploaded documents
* Try editing the document again if loading fails

**Edit modal problems:**

* Wait for content loading to complete before making changes
* Check that the assistant is properly published
* Verify you have editing permissions

## Next Steps

<CardGroup>
  <Card title="Actions & Tools" icon="wrench" href="/advanced-editor/actions-tools">
    Configure integrations and external tools
  </Card>

  <Card title="Playground" icon="gamepad" href="/advanced-editor/playground">
    Test your agent with the knowledge base
  </Card>
</CardGroup>


# Output Settings
Source: https://docs.meetzy.es/advanced-editor/output-settings

Configure what data your agent extracts from each call

## What are Output Settings?

Output Settings define what information your agent should extract and capture from each conversation. This data is then available via webhooks or API for use in your CRM, analytics, or other systems.

<Info>
  Well-configured output settings turn every call into structured, actionable data.
</Info>

## Success Evaluation

Beyond extracting custom data points, you can configure your agent to automatically evaluate whether each call achieved its intended goal. This is done through the **Success** toggle and its associated **Success Criteria** field.

### The Success Toggle

<Info>
  When enabled, the AI evaluates every completed call against your defined criteria and returns a `success` boolean (`true`/`false`) alongside your other output data.
</Info>

<Steps>
  <Step title="Enable Success Tracking">
    Toggle on **Success** in the Output Settings panel to activate call outcome evaluation.
  </Step>

  <Step title="Define Success Criteria">
    Once enabled, a **Success Criteria** text field appears. Describe in natural language what makes a call successful for your specific use case.
  </Step>

  <Step title="AI Evaluates Automatically">
    After each call, the AI reads the full conversation and determines whether the criteria were met.
  </Step>

  <Step title="Access the Result">
    The resulting `success` field is included in your webhook payload and API responses, ready for reporting or automation.
  </Step>
</Steps>

### Writing Effective Success Criteria

The **Success Criteria** field accepts free-form natural language, so you can be as specific as your use case requires. The clearer the description, the more accurate the AI's evaluation will be.

| Agent Type        | Example Success Criteria                                                                                          |
| ----------------- | ----------------------------------------------------------------------------------------------------------------- |
| **Sales Agent**   | "The call is successful if the customer agrees to schedule a demo or provides a budget range."                    |
| **Support Agent** | "The call is successful if the customer's issue was resolved during the conversation without needing escalation." |
| **Booking Agent** | "The call is successful if an appointment date and time were confirmed with the customer."                        |
| **Survey Agent**  | "The call is successful if the customer completed all survey questions."                                          |

<Tip>
  Write success criteria the same way you would explain it to a new team member — describe the specific outcome you're looking for, not just the topic of the call.
</Tip>

<Warning>
  Vague success criteria (e.g. "the call went well") produce inconsistent evaluations. Always tie success to a concrete, observable outcome in the conversation.
</Warning>

## AI Copilot Smart Suggestions

The AI Copilot provides intelligent assistance for configuring output settings with advanced change management capabilities. It analyzes your agent's purpose and conversation goals to recommend optimal data extraction fields.

### How AI Suggestions Work

<Steps>
  <Step title="Analyze Agent Context">
    The AI examines your agent's role, system prompt, and knowledge base to understand its purpose.
  </Step>

  <Step title="Generate Smart Suggestions">
    Based on the analysis, it suggests relevant output parameters and enrichment settings tailored to your use case.
  </Step>

  <Step title="Review Granular Changes">
    You can review, modify, or selectively accept/reject individual parameter suggestions with one click.
  </Step>

  <Step title="Apply Changes">
    Accept all suggestions at once or choose specific ones that fit your needs.
  </Step>
</Steps>

<Info>
  AI suggestions adapt to different agent types - sales agents get lead qualification fields, support agents get issue tracking fields, and scheduling agents get appointment-related outputs.
</Info>

### Example AI Suggestions by Agent Type

| Agent Type        | Suggested Parameters                                                         | Description                            |
| ----------------- | ---------------------------------------------------------------------------- | -------------------------------------- |
| **Sales Agent**   | `lead_score`, `budget_range`, `decision_timeline`, `pain_points`             | Qualify leads and track sales progress |
| **Support Agent** | `issue_type`, `resolution_status`, `satisfaction_score`, `follow_up_needed`  | Track support tickets and outcomes     |
| **Booking Agent** | `appointment_date`, `service_type`, `confirmation_sent`, `preferred_time`    | Manage appointment scheduling          |
| **Survey Agent**  | `survey_responses`, `feedback_rating`, `follow_up_needed`, `completion_rate` | Collect structured feedback            |

## Copilot Change Management

The AI Copilot features an advanced change management system that allows you to review and control all modifications before they're applied. Pending changes are surfaced through dedicated banners so you always know what's about to change before it happens.

### Pending Changes Banners

<Info>
  There are two independent pending changes banners: one for **Output Parameters** and one for **Output Enrichment**. Each tracks its own set of suggestions, so you can review and act on them separately.
</Info>

<Tabs>
  <Tab title="Output Parameters Banner">
    Appears above the Output Parameters section whenever the Copilot proposes adding, removing, or modifying custom parameters.

    * Shows a count badge (e.g. "3 parameter changes")
    * Lists each pending change with an **Add**, **Remove**, or **Modify** label
    * Displays the parameter name and data type for quick scanning
    * Includes **Accept** and **Reject** buttons to apply or dismiss all changes in the batch
  </Tab>

  <Tab title="Output Enrichment Banner">
    Appears above the Output Enrichment section whenever the Copilot proposes enabling or disabling enrichment fields (transcription, summary, evaluations, etc.).

    * Shows a count badge for pending enrichment changes
    * Lists each affected field with its recommended state
    * Includes **Accept** and **Reject** buttons scoped only to enrichment suggestions
  </Tab>
</Tabs>

<Tip>
  Because the two banners are independent, you can accept the AI's Output Parameters suggestions while still reviewing the Output Enrichment recommendations at your own pace (or vice versa).
</Tip>

### Managing Pending Changes

<Tabs>
  <Tab title="Review Changes">
    Each pending change shows:

    * **Operation Type**: Add, Remove, or Modify
    * **Parameter Name**: The field being changed
    * **Data Type**: String, number, or boolean
    * **Description**: What the parameter extracts
  </Tab>

  <Tab title="Granular Control">
    You can:

    * Accept all changes at once via the banner's **Accept** button
    * Reject all changes via the banner's **Reject** button
    * Accept/reject individual parameters
    * Modify suggestions before accepting
  </Tab>

  <Tab title="Change Types">
    The system tracks:

    * **Add**: New parameter suggestions
    * **Remove**: Parameters to be deleted
    * **Modify**: Changes to existing parameters
  </Tab>
</Tabs>

### Change Operation Labels

| Operation  | Color | Description                      |
| ---------- | ----- | -------------------------------- |
| **Add**    | Green | New parameter to be created      |
| **Remove** | Red   | Existing parameter to be deleted |
| **Modify** | Blue  | Changes to existing parameter    |

<Warning>
  Review all pending changes carefully before accepting. Once applied via the **Accept** button, changes will update your agent's output configuration immediately. Use **Reject** to discard the suggestions without modifying your current setup.
</Warning>

## Output Parameters with Templates

Output Parameters are custom fields that your agent extracts from the conversation, now with enhanced template support.

### Parameter Templates System

The new template system provides quick access to common parameter configurations:

<Steps>
  <Step title="Access Templates">
    Click the "Templates" button in Output Settings to see available parameter templates.
  </Step>

  <Step title="Browse Categories">
    Templates are organized by common use cases like customer information, lead qualification, and UTM tracking.
  </Step>

  <Step title="Add Templates">
    Click "Add" next to any template to include it in your output parameters.
  </Step>

  <Step title="Customize After Adding">
    Modify template parameters after adding them to match your specific needs.
  </Step>
</Steps>

### Available Parameter Templates

<Tabs>
  <Tab title="Customer Information">
    * `firstname`, `lastname` - Customer name fields
    * `email`, `phone` - Contact information
    * `company`, `position` - Business details
    * `address`, `city`, `state`, `zip` - Location data
  </Tab>

  <Tab title="Lead Qualification">
    * `lead_score` (number) - Qualification score 1-10
    * `lead_status` (string) - New, Qualified, Disqualified
    * `budget` (number) - Mentioned budget amount
    * `decision_maker` (boolean) - Is decision maker
    * `pain_points` (string) - Customer challenges
  </Tab>

  <Tab title="UTM Tracking">
    * `utm_source`, `utm_medium` - Traffic source tracking
    * `utm_campaign`, `utm_content` - Campaign tracking
    * `utm_term` - Keyword tracking
  </Tab>
</Tabs>

### Creating Custom Parameters

<Steps>
  <Step title="Add Parameter">
    Click "Add Parameter" to create a new custom field.
  </Step>

  <Step title="Configure Properties">
    Set name, type, and description for accurate extraction.
  </Step>

  <Step title="Use AI Assistance">
    Let the Copilot suggest optimal parameters based on your agent's purpose.
  </Step>

  <Step title="Test Extraction">
    Use the Playground to verify the parameter extracts correctly.
  </Step>
</Steps>

### Parameter Properties

| Property    | Description                  | Example                                 |
| ----------- | ---------------------------- | --------------------------------------- |
| Name        | Field identifier (no spaces) | `customer_interest`                     |
| Type        | Data type                    | string, number, boolean                 |
| Description | Clear extraction instruction | "Product the customer is interested in" |

## Advanced Text Mode with Syntax Highlighting

The enhanced text mode now includes full syntax highlighting and error detection for JSON editing.

### Text Mode Features

<Info>
  Text Mode now provides a rich editing experience with syntax highlighting, bracket matching, and real-time error detection for JSON parameter configuration.
</Info>

### Using Text Mode

<Steps>
  <Step title="Switch to Text Mode">
    Toggle the "Text" switch to enter advanced JSON editing mode.
  </Step>

  <Step title="Edit with Syntax Highlighting">
    Benefit from color-coded JSON syntax and automatic formatting.
  </Step>

  <Step title="Real-time Error Detection">
    See validation errors immediately as you type.
  </Step>

  <Step title="Auto-format">
    JSON is automatically formatted for readability.
  </Step>
</Steps>

### Example JSON Configuration

```json theme={null}
[
  {
    "name": "customer_interest",
    "type": "string",
    "description": "Product or service the customer is interested in"
  },
  {
    "name": "appointment_scheduled",
    "type": "boolean", 
    "description": "Whether an appointment was successfully scheduled"
  },
  {
    "name": "satisfaction_score",
    "type": "number",
    "description": "Customer satisfaction rating from 1-5"
  }
]
```

<Tip>
  Text Mode is ideal for copying configurations between agents, bulk editing parameters, or importing from external sources.
</Tip>

## AI-Enhanced Output Enrichment

The AI Copilot now provides intelligent enrichment recommendations based on your agent's specific use case.

### Smart Enrichment Suggestions

<Info>
  The AI analyzes your agent's purpose, industry, and conversation goals to recommend which enrichment fields will provide the most value for your downstream processing needs.
</Info>

### Enrichment Change Management

Just like output parameters, enrichment changes are managed through their own dedicated pending changes banner:

* Review suggested enrichment configurations
* See which fields the AI recommends enabling/disabling
* Accept or reject enrichment suggestions with the banner's **Accept**/**Reject** buttons, independently from parameter changes

### Available Enrichment Fields

| Field                | Description                                                     | AI Suggests For                                             |
| -------------------- | --------------------------------------------------------------- | ----------------------------------------------------------- |
| 📝 **Transcription** | Full conversation transcript                                    | Quality assurance, compliance, training agents              |
| 📋 **Summary**       | AI-generated call summary                                       | Management reporting, CRM integration, executive dashboards |
| ✅ **Success Status** | Whether call achieved its goal (based on your Success Criteria) | Performance tracking, conversion optimization               |
| 🆔 **Call ID**       | Unique call identifier                                          | System integration, debugging, support tickets              |
| 🤖 **Agent ID**      | Playbook/agent identifier                                       | Multi-agent tracking, performance analytics                 |
| ⏱️ **Duration**      | Call length in seconds                                          | Performance metrics, billing, efficiency analysis           |
| 📅 **Created**       | Call start timestamp                                            | Scheduling analysis, trend reporting                        |
| 📊 **Call Status**   | Final call status (completed, failed, etc.)                     | Operational monitoring, success rate tracking               |
| ✅ **Evaluations**    | Results of evaluation tests                                     | Quality assurance, agent training, compliance               |
| 🏷️ **Call Type**    | Inbound/outbound classification                                 | Campaign analysis, channel performance                      |

<Info>
  The **Success Status** enrichment field only returns meaningful data once you've enabled the **Success** toggle and defined your **Success Criteria**. Without it, this field will remain empty.
</Info>

### AI Enrichment Recommendations by Use Case

<Tabs>
  <Tab title="Quality Assurance">
    **Recommended Fields:**

    * Transcription (compliance review)
    * Evaluations (quality scoring)
    * Duration (efficiency tracking)
    * Agent ID (performance monitoring)
  </Tab>

  <Tab title="Sales Operations">
    **Recommended Fields:**

    * Success Status (conversion tracking)
    * Summary (deal progress)
    * Call ID (CRM integration)
    * Duration (sales cycle analysis)
  </Tab>

  <Tab title="Customer Support">
    **Recommended Fields:**

    * Transcription (issue documentation)
    * Summary (resolution tracking)
    * Evaluations (satisfaction scoring)
    * Call Status (operational metrics)
  </Tab>
</Tabs>

<Warning>
  Enable only the enrichment fields you need to keep webhook payloads manageable and reduce processing overhead. The AI helps optimize this selection.
</Warning>

## Webhook Configuration

Send call results to your external systems automatically with enhanced security and reliability features.

### Setting Up Webhooks

<Steps>
  <Step title="Enable Webhook">
    Toggle on the webhook option in Output Settings.
  </Step>

  <Step title="Configure Endpoint">
    Enter your webhook URL and select the HTTP method.
  </Step>

  <Step title="Add Authentication">
    Configure secure authentication headers (Bearer tokens, API keys).
  </Step>

  <Step title="Test Connection">
    Use the built-in test feature to verify your webhook receives data correctly.
  </Step>
</Steps>

### Enhanced Webhook Features

| Feature               | Description                                | Benefit              |
| --------------------- | ------------------------------------------ | -------------------- |
| **Retry Logic**       | Automatic retries for failed webhook calls | Improved reliability |
| **Authentication**    | Secure header-based authentication         | Enhanced security    |
| **Payload Filtering** | Send only enabled enrichment fields        | Reduced bandwidth    |
| **Error Logging**     | Detailed webhook delivery logs             | Better debugging     |

### Example Enhanced Webhook Payload

```json theme={null}
{
  "call_id": "call_abc123",
  "agent_id": "agent_xyz789", 
  "created_at": "2024-01-15T10:30:00Z",
  "duration": 245,
  "status": "completed",
  "success": true,
  "call_type": "outbound",
  "output_params": {
    "customer_interest": "Premium Plan",
    "appointment_scheduled": true,
    "preferred_date": "2024-01-20",
    "satisfaction_score": 5,
    "lead_score": 8,
    "decision_maker": true,
    "pain_points": "Current solution is too expensive",
    "utm_source": "google",
    "utm_campaign": "premium_plan_q1"
  },
  "transcription": "Complete conversation transcript...",
  "summary": "Customer inquired about Premium Plan, qualified as hot lead, appointment scheduled.",
  "evaluations": {
    "greeting_quality": 4.8,
    "needs_assessment": 4.5,
    "closing_effectiveness": 4.2
  }
}
```

<Info>
  The `success` field in the payload above reflects the result of the AI's evaluation against your configured **Success Criteria**. It will be `null` or omitted if the **Success** toggle is disabled.
</Info>

## Best Practices for AI-Enhanced Output Configuration

<Warning>
  Write clear, specific descriptions for output parameters and success criteria. The AI uses these descriptions to understand what to extract and how to judge outcomes. Clear descriptions lead to more accurate data extraction and better AI suggestions.
</Warning>

### Optimization Guidelines

1. **Leverage AI Suggestions** - Start with AI-generated parameters for your agent type
2. **Use Template Library** - Quickly add common parameters from the template system
3. **Define Success Criteria Early** - Enable Success tracking and describe your desired outcome clearly from the start
4. **Review Pending Changes** - Always review AI suggestions in both the Parameters and Enrichment banners before accepting
5. **Be Specific in Descriptions** - "Customer's budget range in USD" vs "budget"
6. **Choose Appropriate Types** - Boolean for yes/no, number for quantities, string for text
7. **Optimize Enrichment** - Enable only necessary enrichment fields based on AI recommendations
8. **Test in Playground** - Verify parameter extraction and success evaluation accuracy before deploying
9. **Secure Webhooks** - Use authentication for sensitive data transmission
10. **Monitor Performance** - Review extraction accuracy and adjust descriptions as needed

### AI Suggestion Acceptance Strategy

<Steps>
  <Step title="Review Agent Context">
    Ensure the AI correctly understands your agent's purpose and industry.
  </Step>

  <Step title="Evaluate Suggestions">
    Review each suggested parameter for relevance to your specific use case.
  </Step>

  <Step title="Customize Descriptions">
    Modify AI-generated descriptions to match your exact requirements.
  </Step>

  <Step title="Test Integration">
    Verify that suggested parameters work with your downstream systems.
  </Step>
</Steps>

## Smart Output Patterns by Industry

### AI-Optimized Sales Call Parameters

```json theme={null}
{
  "lead_qualification_score": "number",
  "interested_product": "string",
  "budget_range": "string", 
  "decision_timeline": "string",
  "pain_points": "string",
  "competitor_mentioned": "string",
  "next_steps": "string",
  "follow_up_needed": "boolean",
  "decision_maker_present": "boolean",
  "company_size": "string",
  "current_solution": "string",
  "implementation_timeline": "string"
}
```

### AI-Optimized Customer Support Parameters

```json theme={null}
{
  "issue_category": "string",
  "issue_severity": "string",
  "product_affected": "string",
  "resolution_provided": "boolean",
  "customer_satisfaction": "number",
  "escalation_needed": "boolean", 
  "resolution_time": "number",
  "follow_up_required": "boolean",
  "root_cause": "string",
  "solution_type": "string",
  "knowledge_base_used": "string"
}
```

### AI-Optimized Healthcare Scheduling Parameters

```json theme={null}
{
  "appointment_date": "string",
  "appointment_time": "string",
  "service_type": "string",
  "provider_requested": "string",
  "insurance_verified": "boolean",
  "symptoms_mentioned": "string",
  "urgency_level": "string",
  "confirmation_method": "string",
  "reminder_preference": "string",
  "special_accommodations": "string"
}
```

## Working with Copilot Suggestions

### Understanding Suggestion Context

When the AI Copilot suggests output parameters, you'll see:

* **Parameter Name** - Optimized field name following naming conventions
* **Data Type** - Recommended type (string, number, boolean) based on expected values
* **Description** - Clear, specific extraction instruction
* **Reasoning** - Why this parameter is relevant to your agent and industry
* **Usage Examples** - Sample values to expect

### Suggestion Management Workflow

<Tabs>
  <Tab title="Batch Operations">
    * **Accept All** - Apply all suggestions in the relevant banner (Parameters or Enrichment) simultaneously
    * **Reject All** - Dismiss all pending changes in that banner
    * **Preview Impact** - See how changes affect your configuration
  </Tab>

  <Tab title="Individual Review">
    * **Parameter-by-Parameter** - Review each suggestion individually
    * **Modify Before Accept** - Edit parameter details before applying
    * **Selective Addition** - Choose only relevant suggestions
  </Tab>

  <Tab title="Advanced Management">
    * **Change Tracking** - See full history of modifications
    * **Rollback Options** - Undo recent changes if needed
    * **Template Integration** - Combine suggestions with template parameters
  </Tab>
</Tabs>

### Customizing AI Suggestions

<Info>
  You can modify AI suggestions before accepting them. This allows you to fine-tune parameter names, adjust descriptions for your specific business context, or change data types based on your integration requirements.
</Info>

## Advanced Integration Scenarios

### Multi-Agent Output Coordination

For organizations running multiple specialized agents:

1. **Consistent Naming** - Use AI suggestions to maintain parameter naming consistency
2. **Cross-Agent Fields** - Implement common fields across different agent types
3. **Enrichment Alignment** - Standardize enrichment settings for unified reporting
4. **Unified Success Criteria** - Define consistent success language across similar agent types for comparable reporting

### CRM Integration Optimization

<Steps>
  <Step title="Map CRM Fields">
    Match AI-suggested parameters to your CRM field structure.
  </Step>

  <Step title="Data Type Alignment">
    Ensure parameter types match your CRM field requirements.
  </Step>

  <Step title="Webhook Testing">
    Test webhook integration with sample AI-generated data.
  </Step>

  <Step title="Error Handling">
    Implement proper error handling for parameter mapping issues.
  </Step>
</Steps>

## Next Steps

<CardGroup>
  <Card title="Greetings" icon="hand-wave" href="/advanced-editor/greetings">
    Configure welcome messages with AI assistance
  </Card>

  <Card title="Evaluations" icon="check-double" href="/advanced-editor/evaluations">
    Set up quality tests and scoring
  </Card>
</CardGroup>


# Playground
Source: https://docs.meetzy.es/advanced-editor/playground

Test your AI agent in real-time with AI streaming, voice commands, and enhanced debugging

## What is the Playground?

The Playground is an enhanced testing environment in the Advanced Editor where you can interact with your agent in real-time. It allows you to simulate conversations, test different scenarios, and validate your agent's behavior before deploying to production, all while leveraging AI streaming capabilities and advanced voice commands.

<Info>
  Testing in the Playground saves time and money by catching issues before they affect real customer calls, with enhanced debugging capabilities through AI streaming and real-time statistics.
</Info>

## Enhanced Editor Integration

The Playground now benefits from the Advanced Editor's improved interface, including:

* **AI Streaming**: Real-time AI responses with streaming capabilities for faster feedback
* **Voice Commands**: Use Cmd+Shift+V for hands-free voice recording and testing
* **Statistics Widgets**: Live performance metrics and real-time debugging information
* **View Controls**: Use outline, indented view, and minimap for better code navigation
* **Diff Mode**: Compare changes with inline/split view options when working with versions
* **Focus Mode**: Minimize distractions while testing
* **Collaborative Testing**: Multiple users can test simultaneously with real-time sync

## Accessing the Playground

The Playground is located in the right panel of the Advanced Editor. Click the "Playground" tab to access it.

<Steps>
  <Step title="Navigate to Advanced Editor">
    Open your agent in the Advanced Editor.
  </Step>

  <Step title="Open Playground">
    Click the "Playground" tab in the right panel.
  </Step>

  <Step title="Configure View">
    Use the editor view controls (outline, minimap, focus mode) to optimize your workspace.
  </Step>

  <Step title="Enable AI Streaming">
    Toggle on AI streaming for real-time responses during testing.
  </Step>
</Steps>

## AI Streaming Capabilities

The Playground now supports advanced AI streaming features:

### Real-time Response Streaming

* **Live Generation**: See AI responses being generated in real-time
* **Token-by-Token Display**: Watch the conversation unfold naturally
* **Interruption Support**: Stop generation mid-stream if needed
* **Performance Metrics**: Monitor response time and token usage

### Statistics Widgets

Real-time widgets display crucial metrics during testing:

<Tabs>
  <Tab title="Performance Metrics">
    * **Response Time**: Live latency measurements
    * **Token Usage**: Input/output token consumption
    * **Memory Usage**: Current conversation context size
    * **Streaming Rate**: Tokens per second generation speed
  </Tab>

  <Tab title="Conversation Stats">
    * **Turn Count**: Number of conversation exchanges
    * **Average Response Length**: Mean response size
    * **Function Calls**: Tool invocation frequency
    * **Context Retention**: Memory usage across turns
  </Tab>

  <Tab title="Quality Metrics">
    * **Coherence Score**: Response consistency rating
    * **Relevance Rating**: Context-appropriate responses
    * **Error Rate**: Failed or incomplete responses
    * **User Satisfaction**: Simulated rating based on flow
  </Tab>
</Tabs>

<Warning>
  Statistics widgets update in real-time and may impact performance during intensive testing sessions. Toggle them off for maximum speed.
</Warning>

## Voice Commands Integration

### Voice Recording (Cmd+Shift+V)

The Playground features enhanced voice input capabilities:

<Steps>
  <Step title="Activate Voice Recording">
    Press `Cmd+Shift+V` (Mac) or `Ctrl+Shift+V` (Windows) to start recording.
  </Step>

  <Step title="Speak Your Input">
    The system will capture and transcribe your voice in real-time.
  </Step>

  <Step title="Review Transcription">
    See live transcription with confidence indicators.
  </Step>

  <Step title="Send or Edit">
    Either send directly or edit the transcribed text before submission.
  </Step>
</Steps>

### Voice Command Features

| Feature         | Description                      | Keyboard Shortcut             |
| --------------- | -------------------------------- | ----------------------------- |
| 🎤 Quick Record | Start/stop voice recording       | `Cmd+Shift+V`                 |
| 🔄 Re-record    | Clear and start new recording    | `Cmd+Shift+R`                 |
| ⏸️ Pause/Resume | Pause recording without stopping | `Spacebar` (during recording) |
| 🔊 Playback     | Hear your recorded audio         | `Cmd+Shift+P`                 |
| 📝 Edit Mode    | Switch to text editing           | `Escape`                      |

<Info>
  Voice commands work with multiple languages and automatically adapt to your agent's language settings.
</Info>

## Call Types with Streaming Support

You can simulate two types of calls with enhanced AI streaming:

| Type        | Description                 | AI Features                   | Voice Commands          |
| ----------- | --------------------------- | ----------------------------- | ----------------------- |
| 📞 Inbound  | Customer calling your agent | Real-time greeting generation | Voice input testing     |
| 📤 Outbound | Agent calling a customer    | Streaming introductions       | Voice script validation |

<Tip>
  Use streaming mode to test how your agent handles interruptions and overlapping speech, crucial for real phone conversations.
</Tip>

## Enhanced Testing Interface

### Monaco Editor with AI Integration

The Playground now features advanced editor capabilities:

<Tabs>
  <Tab title="AI Streaming">
    * **Live Code Analysis**: Real-time syntax validation
    * **Smart Suggestions**: AI-powered completion
    * **Error Detection**: Instant feedback on configuration issues
    * **Performance Hints**: Optimization suggestions
  </Tab>

  <Tab title="Voice Integration">
    * **Speech-to-Code**: Convert voice commands to configuration
    * **Audio Annotations**: Voice notes attached to code sections
    * **Pronunciation Testing**: Validate how text sounds when spoken
    * **Language Detection**: Auto-detect optimal voice settings
  </Tab>

  <Tab title="Collaborative Features">
    * **Live Cursors**: See other users' editing positions
    * **Shared Voice Sessions**: Collaborative voice testing
    * **Real-time Comments**: Voice and text annotations
    * **Change Broadcasting**: Instant updates across sessions
  </Tab>
</Tabs>

## Advanced Message Input

### Text Input with AI Assistance

Type messages with enhanced AI support:

* **Auto-completion**: Smart suggestions based on context
* **Grammar Check**: Real-time language validation
* **Intent Detection**: Preview how the agent will interpret your input
* **Template Suggestions**: Common testing phrases

### Enhanced Voice Input

Click the microphone icon or use `Cmd+Shift+V` for advanced voice features:

<Steps>
  <Step title="Start Voice Recording">
    Activate with keyboard shortcut or button click.
  </Step>

  <Step title="Real-time Transcription">
    See words appear as you speak with confidence indicators.
  </Step>

  <Step title="Language Auto-Detection">
    System automatically detects and switches languages.
  </Step>

  <Step title="Voice Enhancement">
    AI cleans up transcription for better testing accuracy.
  </Step>

  <Step title="Agent Processing">
    Watch streaming response generation in real-time.
  </Step>
</Steps>

## Real-time Conversation Visualization

The enhanced conversation view displays:

| Element          | Description                              | Streaming Features           |
| ---------------- | ---------------------------------------- | ---------------------------- |
| 👤 User Message  | Your inputs with typing indicators       | Live transcription display   |
| 🤖 Agent Message | Streaming AI responses                   | Token-by-token generation    |
| 🎙️ Voice Input  | Audio waveforms and transcription        | Real-time speech recognition |
| 🔧 Tool Call     | Function calls with streaming parameters | Live parameter generation    |
| ✅ Tool Result    | Response with streaming data             | Progressive result display   |

### Audio Playback with AI Enhancement

For each agent response, you can:

* ▶️ Play streaming audio responses
* ⏸️ Pause/resume during playback
* 🔁 Replay with different voice settings
* 📊 Compare audio quality across versions
* 🎛️ Adjust playback speed and tone
* 📈 View audio analytics and clarity metrics

## Statistics Dashboard

### Live Performance Monitoring

The playground features comprehensive real-time statistics:

<AccordionGroup>
  <Accordion title="Response Metrics" icon="chart-line">
    * **Latency Tracking**: Response time distribution
    * **Token Efficiency**: Input/output ratio analysis
    * **Streaming Performance**: Generation speed metrics
    * **Error Rates**: Failed response tracking
  </Accordion>

  <Accordion title="Voice Analytics" icon="microphone">
    * **Recognition Accuracy**: Transcription confidence scores
    * **Audio Quality**: Input clarity measurements
    * **Response Clarity**: Generated audio quality metrics
    * **Language Detection**: Confidence in language identification
  </Accordion>

  <Accordion title="Conversation Flow" icon="comments">
    * **Turn Analysis**: Conversation rhythm tracking
    * **Context Retention**: Memory usage over time
    * **Topic Coherence**: Subject consistency scoring
    * **User Engagement**: Interaction quality metrics
  </Accordion>

  <Accordion title="Resource Usage" icon="server">
    * **Memory Consumption**: Real-time usage tracking
    * **Processing Load**: System resource monitoring
    * **Network Performance**: Connection quality metrics
    * **Cache Efficiency**: Response caching effectiveness
  </Accordion>
</AccordionGroup>

## Context Variables with AI Enhancement

Test dynamic variables with intelligent assistance:

<Steps>
  <Step title="Open Enhanced Context Panel">
    Access context variables with AI-powered validation.
  </Step>

  <Step title="Smart Value Generation">
    Use AI to generate realistic test values for your variables.
  </Step>

  <Step title="Voice Variable Input">
    Speak variable values using voice commands.
  </Step>

  <Step title="Real-time Validation">
    See instant feedback on variable format and content.
  </Step>
</Steps>

### Enhanced Context Management

```json theme={null}
{
  "customer_name": "John Smith",
  "order_number": "ORD-12345", 
  "is_premium_customer": true,
  "account_balance": 1250.50,
  "_ai_context": {
    "streaming_enabled": true,
    "voice_commands_active": true,
    "statistics_level": "detailed",
    "real_time_metrics": true
  },
  "_voice_settings": {
    "input_language": "auto-detect",
    "transcription_confidence": 0.95,
    "voice_enhancement": true
  }
}
```

## Reliability Testing with AI Insights

Run comprehensive tests with AI-powered analysis:

<Steps>
  <Step title="Configure Advanced Scenario">
    Set up test with AI-generated edge cases.
  </Step>

  <Step title="Enable Streaming Mode">
    Test with real-time response generation.
  </Step>

  <Step title="Set Voice Parameters">
    Configure voice input variations for testing.
  </Step>

  <Step title="Run AI-Enhanced Test">
    Execute with streaming metrics and voice analysis.
  </Step>

  <Step title="Analyze AI Insights">
    Review detailed performance analysis and suggestions.
  </Step>
</Steps>

<Warning>
  AI-enhanced reliability tests consume more resources. Consider running during off-peak hours for best results.
</Warning>

## Voice Command Reference

### Global Commands (Available Anywhere)

| Command           | Action                         | Keyboard Shortcut   |
| ----------------- | ------------------------------ | ------------------- |
| "Start recording" | Begin voice input              | `Cmd+Shift+V`       |
| "Stop recording"  | End voice capture              | `Cmd+Shift+V` again |
| "Clear input"     | Reset current input            | `Cmd+Shift+C`       |
| "Send message"    | Submit current input           | `Cmd+Enter`         |
| "Show statistics" | Toggle stats panel             | `Cmd+Shift+S`       |
| "Focus mode"      | Enter distraction-free testing | `Cmd+Shift+F`       |

### Context-Specific Commands

| Context         | Command            | Result                                 |
| --------------- | ------------------ | -------------------------------------- |
| Voice Recording | "Pause"            | Pause recording without losing content |
| Voice Recording | "Resume"           | Continue recording from pause          |
| Voice Recording | "Delete last word" | Remove last spoken word                |
| Voice Recording | "Start over"       | Clear and restart recording            |
| Testing Session | "Run test"         | Execute current test scenario          |
| Testing Session | "Show metrics"     | Display performance dashboard          |

## Advanced Features

### AI-Powered Test Generation

Let AI create test scenarios for you:

<Steps>
  <Step title="Enable AI Test Generator">
    Activate intelligent scenario creation.
  </Step>

  <Step title="Specify Test Goals">
    Describe what you want to test using voice or text.
  </Step>

  <Step title="Review Generated Scenarios">
    AI creates comprehensive test cases with edge conditions.
  </Step>

  <Step title="Execute with Streaming">
    Run tests with real-time monitoring and voice interaction.
  </Step>
</Steps>

### Collaborative Voice Testing

Multiple team members can test together:

* **Shared Voice Sessions**: Multiple users can provide voice input
* **Real-time Transcription Sharing**: See all participants' input
* **Collective Statistics**: Combined performance metrics
* **Voice Queue Management**: Organized turn-taking for voice input

### Intelligent Error Detection

AI continuously monitors for issues:

* **Response Quality**: Automatic coherence checking
* **Performance Degradation**: Real-time slowdown detection
* **Voice Recognition Errors**: Transcription accuracy monitoring
* **Context Loss**: Memory management issue detection

## Best Practices for AI Streaming

<AccordionGroup>
  <Accordion title="Optimize Streaming Performance" icon="zap">
    * Enable streaming only when needed for faster basic testing
    * Use statistics widgets selectively to avoid performance impact
    * Close unnecessary browser tabs during intensive testing
    * Test with realistic network conditions
  </Accordion>

  <Accordion title="Voice Command Efficiency" icon="microphone">
    * Speak clearly and at moderate pace for better recognition
    * Use consistent terminology for reliable command recognition
    * Test in quiet environments for optimal transcription
    * Configure microphone permissions before testing sessions
  </Accordion>

  <Accordion title="Collaborative Testing" icon="users">
    * Establish voice command protocols with team members
    * Use shared statistics to identify performance trends
    * Coordinate testing schedules to avoid resource conflicts
    * Document findings using voice annotations
  </Accordion>

  <Accordion title="Resource Management" icon="server">
    * Monitor system resources during extended testing
    * Use batch testing for high-volume scenarios
    * Cache frequently tested scenarios for faster access
    * Clean up test data regularly to maintain performance
  </Accordion>
</AccordionGroup>

## Troubleshooting Enhanced Features

| Issue                     | Solution                                       | AI Feature                            |
| ------------------------- | ---------------------------------------------- | ------------------------------------- |
| Voice not recognized      | Check microphone permissions and speak clearly | Use voice enhancement settings        |
| Slow streaming            | Disable unnecessary statistics widgets         | Enable performance optimization mode  |
| Statistics not updating   | Refresh statistics panel or restart testing    | AI will auto-detect and suggest fixes |
| Collaborative sync issues | Check network connection and user permissions  | Real-time conflict resolution         |
| Memory usage high         | Clear conversation history and restart session | AI memory management suggestions      |

## Integration with Advanced Systems

### AI Streaming Integration

* **Real-time Sync**: Streaming reflects current version state instantly
* **Change Detection**: Visual indicators for modifications during streaming
* **Performance Optimization**: AI adjusts streaming rate based on system performance
* **Quality Assurance**: Continuous monitoring of stream quality

### Voice Command Integration

* **Cross-Platform Support**: Works across different operating systems
* **Language Adaptation**: Automatically adapts to user's preferred language
* **Context Awareness**: Commands change based on current testing phase
* **Learning System**: AI learns from your voice patterns for better recognition

## Next Steps

<CardGroup>
  <Card title="AI Copilot Integration" icon="robot" href="/advanced-editor/copilot">
    Get AI assistance with enhanced prompt analysis and voice-powered coding
  </Card>

  <Card title="Statistics & Analytics" icon="chart-bar" href="/advanced-editor/analytics">
    Deep dive into performance metrics and AI-powered insights
  </Card>

  <Card title="Voice Configuration" icon="microphone" href="/advanced-editor/voice-settings">
    Configure advanced voice commands and speech recognition settings
  </Card>

  <Card title="Collaborative Features" icon="users" href="/advanced-editor/collaboration">
    Set up team testing sessions with shared voice and streaming capabilities
  </Card>
</CardGroup>


# Tests
Source: https://docs.meetzy.es/advanced-editor/tests

Create and run automated regression tests to ensure your AI agent quality

## What are Tests?

Tests are automated regression tests that validate your agent's behavior. Unlike **Evaluations** (which run after every real call), Tests run on-demand in the Advanced Editor to verify your agent responds correctly to specific scenarios before you deploy changes.

<Info>
  Run tests after modifying prompts, knowledge base, or actions to catch regressions before they affect real customers.
</Info>

## Accessing Tests

Tests are located in the right panel of the Advanced Editor. Click the **Tests** tab (clipboard icon) to access them.

## Tests vs Evaluations

| Feature       | Tests                                       | Evaluations                       |
| ------------- | ------------------------------------------- | --------------------------------- |
| When they run | On-demand in the editor                     | After every real call             |
| Purpose       | Validate specific scenarios before deploy   | Measure call quality over time    |
| Setup         | Define conversation flow + success criteria | Define yes/no questions           |
| Output        | Pass/fail per test                          | Pass/fail per evaluation per call |

## Test Interface Overview

The Tests panel features a modern tabbed interface with two main sections:

<Tabs>
  <Tab title="Tests">
    Your test suite with filtering, creation, and execution controls
  </Tab>

  <Tab title="History">
    Past test runs with detailed results and trend analysis
  </Tab>
</Tabs>

### Header Actions

The header includes quick access to:

* **Context Variables** (`</>`) - Set dynamic variables for test execution
* **Simulate** (💬) - Run one-off conversation simulations
* **Batch Controls** - Run tests multiple times for reliability testing

## Test Types

### Scenario Tests

A **Scenario** test defines a fixed conversation flow. You specify the exact messages (user and agent) and a success condition. The system evaluates whether the agent's response meets your criteria.

| Property          | Description                       | Example                                           |
| ----------------- | --------------------------------- | ------------------------------------------------- |
| Chat history      | Alternating user/agent messages   | User: "Hola" → Agent: "Buenos días..."            |
| Success condition | What the agent should achieve     | "El agente responde de manera profesional y útil" |
| Success examples  | Sample responses that should pass | "Buenos días, ¿en qué puedo ayudarle?"            |
| Failure examples  | Sample responses that should fail | "No tengo idea"                                   |

### Simulation Tests

A **Simulation** test uses an AI to simulate a user. You define a persona, goal, and first message. The system runs a full conversation and evaluates the outcome.

| Property      | Description                   | Example                                 |
| ------------- | ----------------------------- | --------------------------------------- |
| First message | How the simulated user starts | "Hola, buenos días"                     |
| Persona       | Simulated user's profile      | "Cliente interesado en información"     |
| Goal          | What the simulated user wants | "Obtener información sobre el servicio" |

## Creating a Test

<Steps>
  <Step title="Open Tests Panel">
    Click the Tests tab in the right panel and ensure you're on the **Tests** tab (not History).
  </Step>

  <Step title="Click New">
    Click the "New" button to start the creation wizard.
  </Step>

  <Step title="Step 1 - Name & Type">
    Enter a descriptive name and choose Scenario or Simulation.
  </Step>

  <Step title="Step 2 - Conversation">
    For Scenario: Add the conversation flow (user/agent messages). For Simulation: Set first message, persona, and goal.
  </Step>

  <Step title="Step 3 - Criteria">
    Define the success condition and add success/failure examples to guide the evaluation.
  </Step>

  <Step title="Save">
    The test is created and appears in the list.
  </Step>
</Steps>

## Context Variables

Tests can use **dynamic variables** that replace placeholders in your agent's prompts during test execution. Access the Context Variables panel by clicking the **`</>`** icon in the header.

### Built-in Variables

| Variable              | Purpose                      | Example                                              |
| --------------------- | ---------------------------- | ---------------------------------------------------- |
| `fecha_y_hora_actual` | Current date/time in Spanish | "Hoy es Miércoles 12 de Febrero de 2026 a las 14:30" |

### Custom Variables

You can also set values for your agent's input parameters, which will be injected during test execution.

<Steps>
  <Step title="Open Context Variables">
    Click the **`</>`** icon in the Tests header. A badge shows how many variables have values set.
  </Step>

  <Step title="Set Built-in Variables">
    Click the refresh icon next to `fecha_y_hora_actual` to update to the current time.
  </Step>

  <Step title="Set Input Parameters">
    Enter values for any input parameters your agent expects (customer name, order number, etc.).
  </Step>

  <Step title="Run Tests">
    Variables are automatically injected into the agent's context during execution.
  </Step>
</Steps>

<Tip>
  Use `fecha_y_hora_actual` when your agent greets with "Buenos días" or references the current date. Tests run with the values you set, not the live time.
</Tip>

## Batch Testing and Reliability

### Multiple Repetitions

Run tests multiple times to assess reliability and identify intermittent failures:

1. Use the **+/-** controls next to the "Run All" button to set repetition count (1-10)
2. Click **Run All** - the system executes each test the specified number of times
3. View aggregated results in the History tab

<Info>
  Batch testing helps identify consistency issues. If a test passes 8/10 times, you may need to refine your prompts or success criteria.
</Info>

### Batch Progress Tracking

During batch execution:

* Progress indicator shows current run (e.g., "Running 2/5")
* Individual test cards update with interim results
* History tab populates with each completed run

## Running Tests

### Filter Tests

Use filter buttons to focus on specific test states:

* **All** - Show every test
* **Pass** - Only tests that passed last run (with pass count)
* **Fail** - Only tests that failed last run (with fail count)

### Run All Tests

Click **Run All** to execute every test (or all visible tests if filtered). When batch count > 1, button shows "Run x\[count]".

### Run Selected Tests

1. Use checkboxes to select specific tests
2. Click **Run All** (runs only selected when any are checked)

### Run Individual Tests

Click the play icon (▶) on any test card to run just that test.

## Simulate Conversation

The **Simulate** feature lets you run real-time conversation simulations without creating permanent tests:

<Steps>
  <Step title="Open Simulator">
    Click the chat bubble icon (💬) in the Tests header.
  </Step>

  <Step title="Configure Parameters">
    Set first message, persona, goal, and turn limit (default 10).
  </Step>

  <Step title="Start Simulation">
    Click to begin - watch the conversation unfold in real-time with streaming messages.
  </Step>

  <Step title="Review Results">
    See full conversation history and final outcome evaluation.
  </Step>
</Steps>

<Info>
  Simulate is perfect for exploratory testing. Use regular Scenario/Simulation tests when you need repeatable, automated validation.
</Info>

## History Tab

The **History** tab provides comprehensive test run analytics:

### Run Overview

| Column     | Description                               |
| ---------- | ----------------------------------------- |
| Date       | When the run was executed                 |
| Status     | Pass count, fail count, or running status |
| Batch Info | Shows repetition info for batch runs      |
| Actions    | View details, retry failed, retry all     |

### Detailed Results

Click any run to open the detail modal:

* **Individual Test Results** - Pass/fail status with evaluation reasoning
* **Agent Responses** - Full agent output for each test
* **Conversation History** - Complete message flow (for simulation tests)
* **Retry Options** - Re-run failed tests or entire batch

### Trend Analysis

For batch runs with multiple repetitions:

* **Success Rate** - Percentage of tests passing across all repetitions
* **Consistency Metrics** - Identify tests with intermittent failures
* **Performance Patterns** - Spot degradation over time

## Results and Analysis

### Test Cards

Each test card displays:

| Icon | Status  | Meaning                              |
| ---- | ------- | ------------------------------------ |
| ✓    | Pass    | Agent response met success criteria  |
| ✗    | Fail    | Agent response did not meet criteria |
| ⟳    | Running | Test is currently executing          |
| ○    | Pending | Not run yet                          |

### Inline Details

Click a test card to expand inline details:

* **Agent Response** - What the agent actually said
* **Evaluation Reason** - Why it passed or failed
* **Full Conversation** - Complete message history (for simulations)

### Batch Results

For batch runs, test cards show aggregated results:

* **Success Rate** - "8/10 passed" format
* **Latest Result** - Most recent execution status
* **Trend Indicator** - Visual indicator of consistency

## Copilot Integration

Failed tests automatically generate improvement suggestions for Copilot:

<Steps>
  <Step title="Run Tests">
    Execute your test suite and identify failing tests.
  </Step>

  <Step title="Generate Suggestions">
    The system analyzes failures and creates targeted improvement prompts.
  </Step>

  <Step title="Open Copilot">
    Click the generated prompt to open Copilot with context about specific failures.
  </Step>

  <Step title="Apply Improvements">
    Copilot suggests prompt modifications based on test failure analysis.
  </Step>

  <Step title="Re-test">
    Run tests again to validate improvements.
  </Step>
</Steps>

<Warning>
  Copilot integration requires test failures to generate meaningful suggestions. Ensure your success criteria are specific enough to catch real issues.
</Warning>

## Managing Tests

### Edit a Test

1. Click the menu (⋮) on a test card → **Edit**
2. Modify conversation flow, success criteria, or examples
3. Save changes

### Clone a Test

1. Click the menu (⋮) → **Clone**
2. Creates an exact copy for creating variations
3. Useful for testing different personas with same criteria

### Delete a Test

1. Click the menu (⋮) → **Delete**
2. Confirm deletion (irreversible)

### Organize Tests

* **Name Clearly** - Use descriptive names like "Greeting - Professional Tone"
* **Group by Feature** - Create tests for each major capability
* **Version Control** - Clone tests when making significant criteria changes

## Advanced Features

### Real-time Streaming

Simulation tests stream responses in real-time:

* Watch conversations develop naturally
* See agent thinking process
* Identify response delays or issues

### Dynamic Variable Injection

Context variables are injected dynamically:

* Set once, apply to all tests
* Override per-test if needed
* Track variable impact on results

### Historical Comparison

Compare test results across runs:

* Identify regression patterns
* Track improvement over time
* Validate deployment readiness

## Best Practices

<AccordionGroup>
  <Accordion title="Test Critical Paths" icon="route">
    Create tests for your most important flows: greetings, main use case, objections, and compliance.
  </Accordion>

  <Accordion title="Use Success/Failure Examples" icon="scale">
    Provide clear examples so the LLM evaluator understands what "good" and "bad" responses look like.
  </Accordion>

  <Accordion title="Set Context Variables" icon="variable">
    If your agent uses  or input parameters, set them in the context panel before running tests.
  </Accordion>

  <Accordion title="Leverage Batch Testing" icon="repeat">
    Run tests multiple times to identify consistency issues and improve reliability.
  </Accordion>

  <Accordion title="Monitor Trends" icon="chart-line">
    Use the History tab to track test performance over time and catch regressions early.
  </Accordion>

  <Accordion title="Integrate with Development" icon="code">
    Run tests before publishing changes and use Copilot integration for continuous improvement.
  </Accordion>
</AccordionGroup>

## Troubleshooting

### Tests Won't Run

<Steps>
  <Step title="Check Context Variables">
    Ensure required variables like `fecha_y_hora_actual` are set.
  </Step>

  <Step title="Verify Success Criteria">
    Make sure success conditions are specific and measurable.
  </Step>

  <Step title="Review Agent Configuration">
    Confirm your agent has proper prompts and knowledge base setup.
  </Step>
</Steps>

### Inconsistent Results

* **Increase Batch Size** - Run tests 3-5 times to identify patterns
* **Refine Success Criteria** - Make conditions more specific
* **Check Variable Values** - Ensure context variables are set correctly

### Failed Evaluations

* **Review Examples** - Add more success/failure examples
* **Simplify Criteria** - Break complex conditions into multiple tests
* **Use Copilot** - Let AI suggest improvements based on failures

## Next Steps

<CardGroup>
  <Card title="Evaluations" icon="check-double" href="/advanced-editor/evaluations">
    Set up post-call quality metrics
  </Card>

  <Card title="Playground" icon="gamepad" href="/advanced-editor/playground">
    Test your agent interactively
  </Card>

  <Card title="Copilot" icon="robot" href="/advanced-editor/copilot">
    Get AI help improving your prompts
  </Card>

  <Card title="Input Params" icon="sliders" href="/advanced-editor/input-params">
    Configure variables for tests and calls
  </Card>
</CardGroup>


# Version History
Source: https://docs.meetzy.es/advanced-editor/version-history

Git-based version control system for agent configurations with real-time collaboration, conflict resolution, and production deployment workflows

## What is Version History?

Version History provides a complete Git-based version control system for your AI agent configurations, featuring real-time collaboration, intelligent conflict resolution, and safe production deployment workflows. Every modification is tracked with detailed change diffs, collaborative editing locks, and professional commit workflows similar to GitHub or GitLab.

<Info>
  Version History operates like a professional development environment with Git-style workflows. Track changes in real-time, collaborate safely with team members, resolve conflicts intelligently, commit with descriptive messages, and deploy to production with confidence. Your entire development history is preserved with comprehensive audit trails.
</Info>

## Accessing Version History

Version History is seamlessly integrated into the Advanced Editor interface with two main access points:

### Git Status in Header Bar

The editor header displays comprehensive Git status with real-time collaboration indicators:

* **Status badges**: Live indication of working tree state (Clean, Modified, Conflicts, Syncing, Locked)
* **Version display**: Current version number with immediate updates
* **Collaboration indicators**: Active user presence and edit locks
* **Quick action toolbar**: Integrated Check, Pull, and Commit buttons

### Changes Panel

Access the full version control interface through the right panel **"Changes"** tab, providing:

* **Pending Changes**: Review all modifications with detailed diffs before committing
* **History Timeline**: Complete chronological version history with deployment status
* **Commit Workflow**: Professional commit interface with AI-powered message generation

## Git-Style Status System

### Real-Time Status Indicators

The header status badge provides instant feedback on your working tree state:

<Tabs>
  <Tab title="Clean Working Tree">
    **Green Badge: "Clean"**

    * No pending changes, synchronized with remote
    * Version display shows current committed version
    * All changes committed and pushed to remote
    * Ready for new modifications or production deployment
    * Collaboration indicators show any active team members
  </Tab>

  <Tab title="Modified State">
    **Yellow Badge: "Modified (N)"**

    * Shows exact count of changed configuration fields
    * Real-time tracking of specific modified sections
    * Granular change detection at field level
    * Ready to review changes and create commit
    * Auto-save preserves work without committing
  </Tab>

  <Tab title="Conflict State">
    **Red Badge: "Conflicts (N)"**

    * Merge conflicts detected from simultaneous edits
    * Shows count of specific fields requiring resolution
    * Guided conflict resolution workflow available
    * Manual intervention required before committing
    * Visual diff tools highlight conflicting changes
  </Tab>

  <Tab title="Syncing State">
    **Blue Badge: "Syncing..."**

    * Automatic synchronization with remote in progress
    * Background processing of incoming changes
    * Queue status for pending remote updates
    * Non-conflicting changes merged automatically
    * Temporary state during collaboration sync
  </Tab>

  <Tab title="Locked State">
    **Orange Badge: "Locked by X user(s)"**

    * Collaborative editing protection active
    * Shows specific users currently making changes
    * Prevents conflicting simultaneous modifications
    * Automatic lock release after inactivity timeout
    * Real-time collaboration coordination
  </Tab>
</Tabs>

### Integrated Control Toolbar

Quick action buttons seamlessly integrated into the editor header provide professional Git-style workflow control:

<Steps>
  <Step title="Check Button (🔍)">
    **Keyboard Shortcut: ⌘U**

    Manually check for remote updates and potential conflicts. Useful when auto-sync is disabled or when you want to verify remote state before making changes.
  </Step>

  <Step title="Pull Button (⬇️)">
    **Keyboard Shortcut: ⌘⇧P**

    Fetch and merge latest changes from remote repository. Automatically resolves non-conflicting changes and flags conflicts for manual resolution.
  </Step>

  <Step title="Commit Button (✓)">
    **Keyboard Shortcut: ⌘Enter**

    Save current changes with descriptive commit message. Opens professional commit interface with AI-powered message generation and production deployment options.
  </Step>

  <Step title="Push to Production (⬆️)">
    **Keyboard Shortcut: ⌘⇧Enter**

    Deploy committed version directly to production environment. Available after successful commit with immediate live deployment.
  </Step>
</Steps>

## Advanced Change Tracking

### Granular Change Detection

The system provides comprehensive change tracking at multiple levels:

<Tabs>
  <Tab title="Field-Level Precision">
    * **Individual property tracking**: Every configuration field monitored for modifications
    * **Type-aware comparison**: Intelligent comparison based on data types
    * **Nested object support**: Deep tracking of complex configuration structures
    * **Array change detection**: Additions, removals, and reordering in lists
  </Tab>

  <Tab title="Section-Based Organization">
    * **Logical grouping**: Changes categorized by configuration sections
    * **Priority indication**: Critical changes highlighted prominently
    * **Impact assessment**: Visual indication of change significance
    * **Dependency tracking**: Related changes grouped intelligently
  </Tab>

  <Tab title="Real-Time Diffing">
    * **Live comparison**: Instant diff between current and committed state
    * **Syntax highlighting**: Enhanced readability with color-coded changes
    * **Context preservation**: Surrounding unchanged content for clarity
    * **Expandable details**: Drill down into specific change details
  </Tab>

  <Tab title="Intelligent Conflict Detection">
    * **Automatic monitoring**: Continuous checking for conflicting changes
    * **Smart conflict identification**: Detection of semantically conflicting modifications
    * **Resolution suggestions**: AI-powered conflict resolution recommendations
    * **Manual resolution tools**: Visual diff interface for complex conflicts
  </Tab>
</Tabs>

### Enhanced Change Categories

<Warning>
  Different types of changes have varying impact on agent behavior. System prompt modifications directly affect conversation flow, while basic settings changes may only impact metadata. Always test changes thoroughly in Playground before production deployment.
</Warning>

<Tabs>
  <Tab title="Core Configuration">
    **High Impact Changes**

    * **System Prompt**: Primary behavioral instructions and conversation logic
    * **Voice Model**: Speech synthesis and voice characteristics
    * **Response Settings**: Timing, interruption handling, and conversation flow
    * **Language Configuration**: Primary language and accent settings
  </Tab>

  <Tab title="Input/Output Parameters">
    **Workflow Impact Changes**

    * **Parameter Definitions**: New fields, validation rules, and data types
    * **Required Fields**: Changes affecting call completion requirements
    * **Output Webhooks**: Integration endpoints and payload structures
    * **Data Processing**: Transformation rules and formatting logic
  </Tab>

  <Tab title="Actions and Tools">
    **Capability Changes**

    * **Transfer Configuration**: Call routing rules and destination settings
    * **End Call Strategy**: Conversation conclusion logic and criteria
    * **Tool Integrations**: External API connections and capabilities
    * **Workflow Automation**: Business logic and process flows
  </Tab>

  <Tab title="Metadata and Settings">
    **Low Impact Changes**

    * **Agent Name and Description**: Identity and documentation updates
    * **UI Preferences**: Display settings and organizational metadata
    * **Access Permissions**: User roles and visibility settings
    * **Monitoring Configuration**: Analytics and logging preferences
  </Tab>
</Tabs>

## Professional Commit Workflow

### Smart Commit Interface

The commit process follows professional Git workflows with enhanced AI assistance:

<Steps>
  <Step title="Comprehensive Change Review">
    Access the Changes tab to review all pending modifications with expandable diff views, impact assessment, and change categorization for thorough inspection.
  </Step>

  <Step title="Selective Change Staging">
    Use checkboxes to stage specific changes for commit, enabling granular control over what gets included in each version for logical change grouping.
  </Step>

  <Step title="AI-Powered Commit Messages">
    Click "✨ Generate" to create intelligent commit messages based on actual changes made, with context-aware suggestions and best practice formatting.
  </Step>

  <Step title="Production Deployment Toggle">
    Enable "Deploy to Production" to automatically make changes live after successful commit, with safety confirmations and rollback preparation.
  </Step>

  <Step title="Commit Validation and Confirmation">
    Final review dialog shows complete change summary, deployment impact, and confirmation options before creating the new version in timeline.
  </Step>
</Steps>

### Intelligent Commit Message Generation

<Tabs>
  <Tab title="AI-Generated Messages">
    **Context-Aware Generation**

    * **Change Analysis**: AI analyzes actual fields and content modifications
    * **Impact Assessment**: Identifies behavioral changes and their significance
    * **Action-Oriented Language**: Uses clear, descriptive verbs and technical accuracy
    * **Scope Indication**: Includes specific components and areas affected

    **Example Generated Messages:**

    ```
    ✅ Update system prompt to improve customer service tone
    ✅ Add weather lookup capability with OpenWeather integration  
    ✅ Fix webhook timeout configuration for order processing
    ✅ Remove deprecated email notification from end call workflow
    ✅ Optimize response timing for better conversation flow
    ```
  </Tab>

  <Tab title="Best Practice Guidelines">
    **Professional Commit Message Format**

    ```
    Type: Brief description of what changed

    Optional longer explanation of why the change was made,
    what business need it addresses, and any relevant context
    for future developers.

    - Key changes or impacts
    - Testing performed
    - Related tickets or requirements
    ```

    **Recommended Types:**

    * `feat:` New feature or capability
    * `fix:` Bug fix or correction
    * `update:` Modification of existing functionality
    * `config:` Configuration or settings change
    * `docs:` Documentation updates only
  </Tab>

  <Tab title="Common Mistakes to Avoid">
    **Ineffective Commit Messages**

    ```
    ❌ "Fixed stuff"
    ❌ "Updates"
    ❌ "Changed things around"
    ❌ "WIP - work in progress"
    ❌ "Minor tweaks"
    ❌ "Commit before deadline"
    ❌ "Testing"
    ❌ "Temporary changes"
    ```

    **Why These Are Problematic:**

    * No indication of what actually changed
    * Impossible to understand impact without reviewing diffs
    * Makes debugging and rollback decisions difficult
    * Provides no context for team collaboration
  </Tab>
</Tabs>

## Comprehensive Version Timeline

### Professional History Interface

The History timeline provides complete chronological tracking of all agent versions:

<Steps>
  <Step title="Timeline Navigation">
    Browse through complete version history with intuitive chronological layout, visual hierarchy, and comprehensive filtering options.
  </Step>

  <Step title="Rich Version Metadata">
    Each version displays commit message, timestamp, author information, production status, change summary, and performance metrics.
  </Step>

  <Step title="Quick Action Buttons">
    Instant access to Restore Version, Deploy to Production, View Detailed Changes, Compare with Current, and Create Feature Branch.
  </Step>

  <Step title="Detailed Version Inspection">
    Click any version for expanded view with complete diff visualization, metadata analysis, and deployment history.
  </Step>
</Steps>

### Enhanced Version Cards

Each version in the timeline provides comprehensive information for effective version management:

| Element                 | Information                            | Visual Features                                        |
| ----------------------- | -------------------------------------- | ------------------------------------------------------ |
| **Version Number**      | Sequential versioning (v1, v2, v3...)  | Prominent version tag with status indicators           |
| **Commit Message**      | Developer-provided change description  | Primary text with intelligent truncation and expansion |
| **Timestamp**           | Creation time with relative display    | Human-readable relative time (e.g., "2 hours ago")     |
| **Author Information**  | User details with role and permissions | Avatar, name, role badge, and contact information      |
| **Production Status**   | Current deployment and live status     | Prominent "LIVE" badge for active production versions  |
| **Change Summary**      | Quantified modification statistics     | Colored badges showing adds/modifications/deletions    |
| **Performance Metrics** | Success rates and call statistics      | Real-time performance data for deployed versions       |
| **Branch Information**  | Source branch and merge details        | Branch indicators and merge commit relationships       |

## Production Deployment Management

### Safe Production Workflow

<Warning>
  Only the version marked as "Production" handles live customer calls. All other versions remain in development state. Exercise extreme caution when deploying to production, as changes immediately affect customer interactions and business operations.
</Warning>

<Steps>
  <Step title="Pre-Deployment Validation">
    Thoroughly test changes using Playground, run automated regression tests, validate with stakeholders, and ensure rollback plan is ready.
  </Step>

  <Step title="Target Version Selection">
    Navigate the History timeline to identify the specific version for production deployment, review change summary and impact assessment.
  </Step>

  <Step title="Change Impact Review">
    Use "View Changes" to examine detailed modifications, assess business impact, verify compliance requirements, and confirm testing completion.
  </Step>

  <Step title="Production Deployment">
    Click "Mark as Production" to designate version for live traffic, review deployment confirmation dialog showing affected systems.
  </Step>

  <Step title="Deployment Confirmation">
    Final confirmation dialog shows complete impact analysis, affected call volumes, and emergency rollback procedures before deployment.
  </Step>

  <Step title="Post-Deployment Monitoring">
    Monitor live call performance metrics, success rates, error patterns, and be prepared for immediate emergency rollback if issues arise.
  </Step>
</Steps>

### Production Status Management

<Tabs>
  <Tab title="Live Production Indicators">
    * **Prominent "LIVE" badge**: Clearly identifies currently deployed version
    * **Crown icon**: Visual hierarchy indicator for production status
    * **Green border highlighting**: Makes production version stand out in timeline
    * **Real-time metrics**: Live performance statistics and call success rates
    * **Traffic routing status**: Confirmation of active call routing to this version
  </Tab>

  <Tab title="Deployment History Tracking">
    * **Deployment timestamps**: Complete record of when each version went live
    * **Duration tracking**: How long each version served production traffic
    * **Rollback indicators**: Versions that were emergency rolled back with reasons
    * **Performance correlation**: Success metrics tied to specific version deployments
    * **Business impact analysis**: Revenue and customer satisfaction correlations
  </Tab>

  <Tab title="Multi-Environment Support">
    * **Staging environment**: Safe testing environment mirroring production
    * **A/B testing capability**: Gradual rollout to percentage of traffic
    * **Blue-green deployment**: Zero-downtime deployment strategy support
    * **Canary releases**: Limited exposure testing with automatic rollback
    * **Environment synchronization**: Consistent configuration across environments
  </Tab>
</Tabs>

## Real-Time Collaboration Features

### Advanced Collaboration System

Professional team collaboration with conflict prevention and resolution:

<Steps>
  <Step title="Live User Presence">
    Real-time display of all team members currently editing the agent, with activity status and current focus areas for coordination.
  </Step>

  <Step title="Intelligent Conflict Prevention">
    Automatic locking system prevents simultaneous edits to the same configuration sections, with clear lock indicators and timeout management.
  </Step>

  <Step title="Continuous Auto-Sync">
    Background synchronization every 30 seconds checks for remote changes, with smart merging of non-conflicting modifications automatically.
  </Step>

  <Step title="Guided Conflict Resolution">
    When conflicts occur, visual diff interface provides step-by-step resolution workflow with merge options and preview capabilities.
  </Step>
</Steps>

### Collaboration State Management

<Tabs>
  <Tab title="Active Collaboration Indicators">
    * **User avatars**: Live display of currently active editors with status
    * **Section-level locks**: Visual indicators when specific areas are being edited
    * **Real-time sync status**: Header badge showing current collaboration state
    * **Change notifications**: Instant alerts when remote changes are detected
    * **Activity timeline**: Recent collaborative activities and contributions
  </Tab>

  <Tab title="Conflict Resolution Interface">
    * **Visual conflict highlighting**: Clear markers for conflicting configuration sections
    * **Resolution options**: Choose local changes, remote changes, or create custom merge
    * **Preview functionality**: See results of conflict resolution before final commitment
    * **Guided step-by-step process**: Wizard interface for complex conflict scenarios
    * **Backup and recovery**: Automatic backups during conflict resolution process
  </Tab>

  <Tab title="Auto-Sync Intelligence">
    * **Background monitoring**: Continuous polling for remote repository changes
    * **Smart merging algorithms**: Automatic resolution of non-conflicting modifications
    * **Conflict prediction**: Early warning system for potential editing conflicts
    * **Manual override controls**: Force sync options when automatic resolution fails
    * **Network resilience**: Graceful handling of connectivity issues and recovery
  </Tab>
</Tabs>

### Pull and Sync Operations

<Steps>
  <Step title="Automatic Sync Monitoring">
    System continuously monitors remote repository for changes, with configurable polling intervals and immediate notification of conflicts.
  </Step>

  <Step title="Manual Check Operation">
    Use "Check" button (⌘U) to manually verify remote change status, bypass automatic sync, and get immediate status update.
  </Step>

  <Step title="Pull Remote Changes">
    Use "Pull" button (⌘⇧P) to fetch and integrate all remote changes, with automatic conflict detection and resolution guidance.
  </Step>

  <Step title="Conflict Resolution Workflow">
    When conflicts are detected, guided interface helps resolve each conflict with visual diff tools and merge options.
  </Step>

  <Step title="Sync Completion Verification">
    Status automatically updates to "Clean" when local and remote states are fully synchronized and all conflicts resolved.
  </Step>
</Steps>

## Advanced Workflow Patterns

### Emergency Production Hotfix

<Steps>
  <Step title="Critical Issue Identification">
    Rapidly identify urgent production issue requiring immediate configuration change through monitoring alerts or customer reports.
  </Step>

  <Step title="Targeted Quick Fix">
    Make precise, minimal changes to only the specific configuration elements needed to resolve the critical issue.
  </Step>

  <Step title="Rapid Validation Testing">
    Use Playground for immediate verification that the fix resolves the issue without introducing unintended side effects.
  </Step>

  <Step title="Emergency Commit and Deploy">
    Use "Deploy to Production" toggle during commit to push changes live immediately, bypassing normal review cycles.
  </Step>

  <Step title="Live Monitoring and Verification">
    Monitor call analytics and performance metrics in real-time to confirm successful resolution of the production issue.
  </Step>

  <Step title="Post-Incident Documentation">
    Update commit messages with detailed incident context for future reference and create post-mortem documentation.
  </Step>
</Steps>

### Feature Development with Branching

<Steps>
  <Step title="Feature Branch Creation">
    Create new development branch from stable main branch for isolated feature development without affecting production.
  </Step>

  <Step title="Iterative Development Process">
    Make incremental changes with frequent commits, descriptive messages, and regular testing to maintain development momentum.
  </Step>

  <Step title="Comprehensive Testing Protocol">
    Use Playground extensively, run automated regression tests, and validate in staging environments before merge consideration.
  </Step>

  <Step title="Team Code Review Process">
    Collaborative review of changes using integrated diff views, comment system, and approval workflows before integration.
  </Step>

  <Step title="Main Branch Integration">
    Merge completed and fully tested feature back to main development branch with conflict resolution as needed.
  </Step>

  <Step title="Production Deployment Strategy">
    Deploy merged version to production with comprehensive monitoring, gradual rollout, and prepared rollback procedures.
  </Step>
</Steps>

### Emergency Rollback Procedures

<Steps>
  <Step title="Issue Detection and Assessment">
    Quickly identify production issues through automated monitoring, customer reports, or team observations with severity assessment.
  </Step>

  <Step title="Impact Analysis and Decision">
    Evaluate issue severity, customer impact, and business consequences to determine if immediate rollback is necessary.
  </Step>

  <Step title="Last Known Good Version">
    Use History timeline to quickly identify the most recent stable production version with confirmed good performance.
  </Step>

  <Step title="Emergency Version Restore">
    Use "Restore This Version" to immediately revert configuration to last stable state, bypassing normal review processes.
  </Step>

  <Step title="Immediate Production Deployment">
    Mark restored version as Production to stop customer impact immediately and restore service to known good state.
  </Step>

  <Step title="Post-Rollback Analysis">
    After stability is restored, analyze the problematic version to understand root cause and prevent future occurrences.
  </Step>
</Steps>

## Comprehensive Keyboard Shortcuts

### Primary Version Control Actions

| Action                      | Windows/Linux      | macOS     | Description                                                    |
| --------------------------- | ------------------ | --------- | -------------------------------------------------------------- |
| **Open Changes Panel**      | `Ctrl+G`           | `⌘G`      | Switch to Changes tab for reviewing pending modifications      |
| **Check for Updates**       | `Ctrl+U`           | `⌘U`      | Manually check remote repository for new changes and conflicts |
| **Pull Remote Changes**     | `Ctrl+Shift+P`     | `⌘⇧P`     | Fetch and merge latest changes from remote repository          |
| **Open Commit Dialog**      | `Ctrl+Enter`       | `⌘Enter`  | Open professional commit interface for current changes         |
| **Deploy to Production**    | `Ctrl+Shift+Enter` | `⌘⇧Enter` | Push committed version directly to production environment      |
| **Generate Commit Message** | `Ctrl+Shift+G`     | `⌘⇧G`     | AI-generate descriptive commit message based on changes        |

### History and Navigation

| Action                     | Windows/Linux  | macOS | Description                                                       |
| -------------------------- | -------------- | ----- | ----------------------------------------------------------------- |
| **View History Timeline**  | `Ctrl+H`       | `⌘H`  | Switch to History view for complete version timeline              |
| **Quick Rollback**         | `Ctrl+Shift+R` | `⌘⇧R` | Quick access to emergency rollback to last production version     |
| **Compare Versions**       | `Ctrl+D`       | `⌘D`  | Open side-by-side comparison between current and selected version |
| **Search Version History** | `Ctrl+F`       | `⌘F`  | Search version history by commit message, author, or date         |

### Change Review and Editing

| Action                   | Windows/Linux  | macOS | Description                                         |
| ------------------------ | -------------- | ----- | --------------------------------------------------- |
| **Expand All Changes**   | `Ctrl+Shift+E` | `⌘⇧E` | Expand all change sections for comprehensive review |
| **Collapse All Changes** | `Ctrl+Shift+C` | `⌘⇧C` | Collapse all sections for overview and navigation   |
| **Next Change**          | `Ctrl+Down`    | `⌘↓`  | Navigate to next changed configuration section      |
| **Previous Change**      | `Ctrl+Up`      | `⌘↑`  | Navigate to previous changed configuration section  |
| **Toggle Diff View**     | `Alt+D`        | `⌥D`  | Switch between inline and side-by-side diff views   |

## Production Best Practices

<AccordionGroup>
  <Accordion title="Frequent Commits with Clear Messages" icon="pencil">
    **Atomic Commits Strategy**

    Create commits after each logical change rather than bundling multiple unrelated modifications. This creates more granular restore points and makes the timeline easier to navigate for debugging and collaboration.

    * One feature or fix per commit
    * Clear, descriptive commit messages explaining what and why
    * Include business context and requirements references
    * Test each commit individually before pushing
  </Accordion>

  <Accordion title="Comprehensive Testing Before Deployment" icon="beaker">
    **Multi-Layer Testing Approach**

    Always validate changes thoroughly before production deployment to minimize customer impact and business disruption.

    * Use Playground for conversation flow testing
    * Run automated regression test suites
    * Validate in staging environments that mirror production
    * Perform load testing for performance-critical changes
    * Get stakeholder approval for behavioral changes
  </Accordion>

  <Accordion title="Strategic Branching for Complex Work" icon="git-branch">
    **Branch Management Strategy**

    Use feature branches effectively to isolate development work and enable parallel development without conflicts.

    * Create feature branches for major changes or experiments
    * Keep main branch stable and deploy-ready
    * Use descriptive branch names reflecting the work
    * Merge branches only after thorough testing and review
    * Delete branches after successful merge to avoid clutter
  </Accordion>

  <Accordion title="Proactive Production Monitoring" icon="chart-line">
    **Continuous Production Oversight**

    Monitor agent performance continuously after deployments to catch issues early and ensure optimal performance.

    * Set up automated alerts for performance degradation
    * Monitor key metrics: success rate, response time, error patterns
    * Establish baseline metrics before major changes
    * Prepare rollback procedures before each deployment
    * Document monitoring procedures for team consistency
  </Accordion>

  <Accordion title="Collaborative Conflict Prevention" icon="users">
    **Team Coordination Best Practices**

    Prevent conflicts through communication and coordinated development practices.

    * Communicate major changes before starting work
    * Use the collaboration lock system effectively
    * Resolve conflicts promptly when they occur
    * Maintain awareness of team member activities
    * Establish team conventions for commit messages and workflows
  </Accordion>

  <Accordion title="Comprehensive Documentation" icon="document-text">
    **Change Documentation Standards**

    Document not just what changed, but why changes were made and their expected impact.

    * Include business context in commit messages
    * Reference tickets, issues, or requirements driving changes
    * Document testing performed and results achieved
    * Note any special deployment considerations or risks
    * Maintain changelog for major feature releases
  </Accordion>
</AccordionGroup>

## Troubleshooting and Recovery

### Common Scenarios and Solutions

<AccordionGroup>
  <Accordion title="Resolving Merge Conflicts" icon="exclamation-triangle">
    **Step-by-Step Conflict Resolution**

    When conflicts appear, the system provides a guided resolution process:

    1. **Identify Conflicts**: Changes tab highlights conflicting fields in red
    2. **Review Changes**: Use diff interface to understand both local and remote changes
    3. **Choose Resolution**: Select local changes, remote changes, or create custom merge
    4. **Preview Results**: Review the merged result before committing
    5. **Complete Resolution**: Commit the resolved changes to complete the merge

    Each conflict must be resolved individually for complete merge completion.
  </Accordion>

  <Accordion title="Recovering Lost Changes" icon="arrow-uturn-left">
    **Change Recovery Process**

    All changes are preserved in version history - nothing is permanently lost:

    * **Check History Timeline**: Review recent commits for your changes
    * **Search by Author**: Filter history by your username
    * **Search by Message**: Use commit message search to locate changes
    * **Check Uncommitted Changes**: Look in Changes tab for unsaved work
    * **Branch Investigation**: Changes might be in feature branches

    If changes still seem missing, contact support with specific details about when changes were made.
  </Accordion>

  <Accordion title="Emergency Production Response" icon="exclamation-circle">
    **Critical Issue Response Protocol**

    For production emergencies requiring immediate action:

    1. **Assess Impact**: Determine severity and customer impact scope
    2. **Emergency Rollback**: Use quick rollback to last known good version
    3. **Verify Resolution**: Confirm issue is resolved with rollback
    4. **Communicate Status**: Update stakeholders on resolution
    5. **Investigate Root Cause**: Analyze problematic version after stability restored
    6. **Implement Preventive Measures**: Update processes to prevent recurrence

    Document all emergency actions for post-incident analysis.
  </Accordion>

  <Accordion title="Synchronization Issues" icon="arrow-path">
    **Sync Problem Resolution**

    When auto-sync shows persistent errors or fails to complete:

    * **Manual Sync**: Use Check and Pull buttons to force synchronization
    * **Clear Browser Cache**: Refresh authentication and cached data
    * **Network Verification**: Check internet connectivity and firewall settings
    * **Authentication Check**: Verify login status and permissions
    * **Browser Refresh**: Complete page reload to reset sync state

    If problems persist after these steps, contact technical support with error details.
  </Accordion>

  <Accordion title="Collaboration Lock Resolution" icon="lock-closed">
    **Resolving Edit Lock Issues**

    When locked out due to collaborative editing conflicts:

    * **Wait for Timeout**: Locks automatically release after 30 minutes of inactivity
    * **Contact Active User**: Reach out directly to the user holding the lock
    * **Save Work Locally**: Preserve your changes to prevent loss during resolution
    * **Admin Override**: System administrators can manually release stuck locks
    * **Alternative Approach**: Use feature branches to work around active locks

    Document recurring lock issues to improve team coordination processes.
  </Accordion>

  <Accordion title="Performance and Deployment Issues" icon="exclamation-circle">
    **Deployment Problem Response**

    When deployed changes cause performance problems or unexpected behavior:

    1. **Immediate Assessment**: Check call analytics for impact measurement
    2. **Traffic Analysis**: Monitor call volume, success rates, and error patterns
    3. **Gradual Rollout**: If available, limit exposure using A/B testing
    4. **Performance Baselines**: Compare against pre-deployment metrics
    5. **Emergency Rollback**: If severe issues, immediately revert to stable version
    6. **Issue Documentation**: Record problems for future deployment planning

    Always establish performance baselines before significant changes.
  </Accordion>
</AccordionGroup>

## Integration with Advanced Editor

### Seamless Cross-Feature Integration

Version History is deeply integrated with all Advanced Editor features for unified workflow:

<Tabs>
  <Tab title="Monaco Editor Integration">
    * **Real-time change detection**: Every keystroke and modification tracked automatically
    * **Syntax-aware diffing**: Intelligent comparison considering code structure
    * **Auto-save with version tracking**: Continuous preservation without committing
    * **Conflict highlighting**: Visual indicators directly in editor for conflicts
  </Tab>

  <Tab title="AI Copilot Integration">
    * **Suggestion tracking**: AI recommendations tracked as potential changes
    * **Commit message generation**: Copilot analyzes changes for intelligent descriptions
    * **Change impact analysis**: AI assessment of modification significance
    * **Optimization suggestions**: Recommendations for commit organization
  </Tab>

  <Tab title="Playground Integration">
    * **Test result correlation**: Link test outcomes to specific versions
    * **Pre-commit validation**: Test results inform deployment readiness
    * **Performance benchmarking**: Testing data included in commit context
    * **Regression detection**: Automated comparison with previous test results
  </Tab>

  <Tab title="Testing Suite Integration">
    * **Automated test triggers**: Version changes trigger regression test suites
    * **Deployment blocking**: Failed tests prevent production deployment
    * **Test history correlation**: Link test results to specific version deployments
    * **Quality gate enforcement**: Require passing tests before production marking
  </Tab>
</Tabs>

### Workflow Automation

<Steps>
  <Step title="Change Detection Pipeline">
    Monaco Editor modifications automatically flow into version tracking with real-time diff generation and change categorization.
  </Step>

  <Step title="AI-Enhanced Commits">
    Copilot analyzes pending changes to generate contextual commit messages and impact assessments for informed decisions.
  </Step>

  <Step title="Testing Integration">
    Playground test results and automated test suite outcomes are incorporated into deployment decision workflows.
  </Step>

  <Step title="Production Pipeline">
    Successful commits can trigger automated deployment pipelines with safety checks and monitoring integration.
  </Step>
</Steps>

## Next Steps

<CardGroup>
  <Card title="Test Changes in Playground" icon="play" href="/advanced-editor/playground">
    Validate your modifications with comprehensive conversation testing before committing changes to version history
  </Card>

  <Card title="AI Copilot Assistant" icon="sparkles" href="/advanced-editor/copilot">
    Get intelligent assistance with commit messages, change analysis, and optimization recommendations
  </Card>

  <Card title="Automated Testing Suite" icon="clipboard-check" href="/advanced-editor/tests">
    Set up comprehensive regression tests to prevent production issues and ensure quality
  </Card>

  <Card title="Production Call Analytics" icon="chart-bar" href="/monitoring/call-analytics">
    Monitor real-time performance metrics and success rates after deploying configuration changes
  </Card>
</CardGroup>


# Manage Agents
Source: https://docs.meetzy.es/agents/manage-agents

Learn how to view, edit, clone, archive, and manage your voice agents in the redesigned Meetzy Agents dashboard

## Overview

The **Agents** page has been fully redesigned around a **master-detail layout**: a compact agent list on the left and a rich detail panel on the right, so you can browse, inspect, and act on any agent without leaving the page.

<Info>
  This redesign replaces the classic Agents page. A **"Switch back to the classic view"** banner sits at the very top of the screen while both experiences coexist — click it any time to return to the previous layout without losing your place.
</Info>

## What's New in the Agents Page (Agents v2)

The new Agents page (internally called **Agents v2**) bundles several improvements that were previously scattered across different views or unavailable altogether:

* **Latency history per agent** — every agent row shows a compact latency sparkline, and the detail panel's Overview tab includes a full 14-day latency history chart, so you can spot regressions at a glance or dive deeper without leaving the page.
* **Lazy-loaded Phone Health chips** — a colored beacon next to each agent's number reflects its spam/deliverability status (green, red, or gray), with a direct link into the full [Phone Health](/agents/phone-numbers) report. To keep the page fast, this status is fetched on demand rather than for the whole list at once.
* **Sortable agent list** — a dedicated sort selector lets you reorder the list by newest, most calls, or name.
* **Richer filters** — search, company, archived visibility, and a dedicated spam-only toggle now live together in a single filters bar, with a one-click **Clear filters** action.
* **New "Call" modal** — placing a test call (inbound or outbound) now happens from a redesigned modal with country prefix selection, agent version selection for Advanced agents, and a shortcut to edit custom form fields, all without leaving the Agents page.
* **Collapsible demo agents section** — sample/demo agents are grouped separately from your real agents and tucked away by default, keeping the main list focused.
* **A brand-new detail panel** — selecting an agent now opens a tabbed **Overview / Calls / Settings** panel packed with metrics, a 90-day call log, and the full configuration form, all in one place.

<Tip>
  Missing the old page already? The **"Switch back to the classic view"** link at the top of the screen is still there for you — no judgment if you, like Wolverine pining for 1980s Canada, occasionally long for the way things used to be.
</Tip>

<Tabs>
  <Tab title="Agents List (left)">
    A scrollable list of agents with search, company filter, a sort selector, and an **Archived agents** visibility toggle. Selecting an agent loads its details on the right without a full page reload. Each row also surfaces a phone health indicator and a latency mini-chart at a glance. Demo agents appear in their own collapsible section below your real agents.
  </Tab>

  <Tab title="Detail Panel (right)">
    A tabbed panel — **Overview**, **Calls**, and **Settings** — showing everything about the selected agent: performance metrics, latency history, call log, and configuration.
  </Tab>
</Tabs>

## Agent List and Filters

### Search and Company Filter

The filters bar above the list lets admins and multi-company users narrow down the agent list quickly:

| Filter              | Description                                                                                                                  |
| ------------------- | ---------------------------------------------------------------------------------------------------------------------------- |
| **Search**          | Matches agent name, company, or agent ID in real time                                                                        |
| **Company**         | Dropdown with a built-in search box to filter agents by organization; defaults to **"All companies"**                        |
| **Archived agents** | Toggle switch to show or hide archived agents in the list (hidden by default)                                                |
| **Phone health**    | Toggle switch to show only agents whose numbers currently have spam/deliverability issues                                    |
| **Clear filters**   | Appears only when at least one filter is active, resetting search, company, archived visibility, and the phone health filter |

<Tip>
  The company filter dropdown includes its own search box, making it easy to jump straight to a specific organization when you manage dozens of companies.
</Tip>

### Sorting the List

A compact sort selector sits above the agent list, next to the agent count, and lets you reorder the list without affecting any of the active filters:

| Sort option                | Order                                                                                  |
| -------------------------- | -------------------------------------------------------------------------------------- |
| **Newest first** (default) | Most recently created agents at the top                                                |
| **Most calls**             | Agents ranked by call volume in the same 14-day window used for the latency sparklines |
| **Name (A-Z)**             | Alphabetical, ascending                                                                |
| **Name (Z-A)**             | Alphabetical, descending                                                               |

<Info>
  Your sort preference persists across visits, so you don't need to reselect it every time you open the Agents page.
</Info>

### Selecting an Agent

Click any row in the list to open it in the detail panel. The selected row is highlighted with a colored left border, and archived agents display a small **Archived** badge next to their name.

### Phone Health Indicator and Latency Mini-Chart

Every row in the agent list now includes two at-a-glance signals, so you no longer need to open an agent to spot a problem:

* **Phone health beacon**: a small colored dot next to the agent's phone number reflecting the spam/deliverability status of its assigned numbers (green = healthy, red = spam issues detected, gray = status unknown or still loading). Click it to jump straight into the [Phone Health](/agents/phone-numbers) report for that agent.
* **Latency sparkline**: a compact 14-day mini-chart of average AI response latency, rendered inline in the row. It uses the same underlying data as the Overview tab's latency history chart, letting you compare agents' recent performance without opening each one individually.

<Info>
  **Latency history** is fetched in bulk (in small chunks) right after the list loads, so sparklines populate for every visible row a moment after the initial render.

  **Phone health**, however, is loaded **lazily** and per agent — it's a slower check that isn't practical to run for hundreds of agents at once. The beacon shows a neutral "Check" state until you either select that agent (which triggers its individual health check) or turn on the **Phone health** spam-only filter, which loads the status for every currently visible agent so the filter can work correctly.
</Info>

### Collapsible Demo Agents

Sample and demo agents are grouped separately from your production agents, in a dedicated **Demo agents** section below the main list. This section is **collapsed by default** to keep the primary list focused on the agents you actually manage — click its header to expand or collapse it. Demo agents can still be selected, previewed, and called for testing purposes like any other agent, but they're excluded from the main list's filters and sort order.

## Agent Detail Panel

Selecting an agent from the list opens the detail panel on the right, with a header showing the agent's name, ID badge, and company, followed by a set of quick action buttons and three tabs: **Overview**, **Calls**, and **Settings**.

<Info>
  On mobile and narrow screens, the header actions collapse into a single kebab (**⋮**) menu to save space, exposing the same options in a dropdown.
</Info>

### Header Actions

A row of buttons at the top of the panel gives you one-click access to the most common agent workflows, without needing to scroll into the Overview tab:

| Action      | Availability                                   | What it does                                                                     |
| ----------- | ---------------------------------------------- | -------------------------------------------------------------------------------- |
| **Editor**  | Advanced agents only                           | Opens the [Advanced Editor](/agents/advanced-editor) for this agent in a new tab |
| **Trends**  | Advanced agents only                           | Opens the evaluation trends view for this agent in a new tab                     |
| **Clone**   | Admins and users/members with clone permission | Launches the two-step clone assistant described below                            |
| **Archive** | Users with agent create/edit permission        | Launches the archive confirmation flow described below                           |

<Tabs>
  <Tab title="Overview">
    **Performance dashboard** for the agent, including:

    * **Summary cards**: total calls, average call duration, and average AI latency (with a trend arrow comparing recent vs. previous performance, plus a P90 latency reading)
    * **Latency history chart**: a 14-day time series of average AI response latency, helping you spot regressions or improvements over time
    * **Call volume chart**: a 30-day time series of daily call counts
    * **Phone health**: a clickable card linking to the spam/deliverability report for the agent's assigned numbers
  </Tab>

  <Tab title="Calls">
    A sortable table of the agent's calls from the **last 90 days**, including date, origin number, status, duration, and latency. Click any column header to reorder the table by that field. Admin users additionally see quick links per call out to **Twilio** and **ElevenLabs** for deeper debugging. Recordings can be played directly from this tab using the same player used across the Calls section, letting you inspect a specific call without leaving the panel. A **"View all calls"** shortcut opens the main Calls section in a new tab, already filtered down to this agent, for whenever the 90-day snapshot isn't enough.
  </Tab>

  <Tab title="Settings">
    The full configuration form for standard (non-Advanced) agents — the same settings previously available on the classic page, including save/discard change detection and an editing timeout safeguard. Advanced agents don't show this tab since their configuration is managed from the [Advanced Editor](/agents/advanced-editor).
  </Tab>
</Tabs>

<Info>
  The **Avg AI latency** trend arrow compares the average of the most recent half of the visible days against the earlier half — an upward red arrow means latency is increasing, a downward green arrow means it's improving. It only appears once enough daily data points are available.
</Info>

### Summary Cards

The three cards at the top of the Overview tab give you an instant read on agent health:

| Card               | What it shows                                                                                      |
| ------------------ | -------------------------------------------------------------------------------------------------- |
| **Total calls**    | Total number of calls handled by the agent                                                         |
| **Avg duration**   | Average call length, formatted as minutes:seconds                                                  |
| **Avg AI latency** | Average AI response latency in milliseconds, its trend arrow, and the P90 latency value underneath |

### Latency History and Call Volume Charts

Below the summary cards, two charts plot the agent's daily behavior over time:

* **Latency history (14 days)**: average AI latency per day, plotted chronologically to make regressions easy to spot.
* **Call volume (30 days)**: number of calls handled per day, useful for spotting usage spikes or drop-offs.

Both charts share the same rendering component and gracefully show an empty state message when there isn't enough data yet.

### Phone Health Shortcut

The **Phone health** card sits below the charts and links directly to the spam/deliverability report for the numbers assigned to the selected agent — the same report reachable from the row-level beacon in the agent list. Selecting an agent from the list is exactly what triggers its individual phone health check, so this card should always have fresh data by the time you view it.

### Calls Tab in Detail

The Calls tab gives you a focused, agent-scoped view of call history without needing to jump to the main Calls section and filter manually:

* **90-day window**: the table loads calls from the last 90 days by default, sorted by most recent first.
* **Sortable columns**: click any column header (date, status, duration, latency, etc.) to toggle its sort order.
* **Admin-only debugging links**: admin users see extra per-call links out to the underlying **Twilio** call log and the **ElevenLabs** conversation, useful when troubleshooting a specific problem call.
* **Inline recording playback**: click a call to open its recording in the same player component used in the main Calls section, so you can listen to (or download) the audio without leaving the agent's detail panel.
* **View all calls**: opens the full [Calls](/calls) section in a new tab, pre-filtered to this agent's ID, for whenever you need to look further back than 90 days or apply additional filters.

<Tip>
  This shortcut mirrors an option that existed on the classic Agents page — it opens the general Calls list scoped to the agent instead of duplicating advanced filtering inside the detail panel.
</Tip>

### Settings Tab in Detail

For standard (non-Advanced) agents, the Settings tab reproduces the full classic configuration form inside the detail panel:

* **Agent name** and core identification fields
* **Prompts**: the agent's system/behavior prompts, editable inline or via a fullscreen textarea for longer content
* **Outbound options**: call behavior settings applied to outbound calls
* **Evaluations**: manage the list of post-call evaluations run against this agent's conversations, including creating, editing, and deleting evaluation criteria
* **Webhook**: configure the webhook endpoint that receives call events and data for this agent
* **NPS survey**: enable and configure a post-call satisfaction survey
* **Danger zone**: a collapsed, opt-in section exposing sensitive, rarely-touched settings — raw instructions, assigned phone numbers, and call transfer configuration

<Warning>
  The **Danger zone** fields directly affect how the agent behaves in production calls. Double-check raw instruction changes, phone number assignments, and transfer settings before saving.
</Warning>

<Info>
  Settings changes are tracked with a **dirty-check**: the Save button only activates once you've actually modified something, and an **editing timeout** automatically locks the form after **10 minutes of inactivity**, requiring you to reload the latest data before making further changes. This prevents two people from silently overwriting each other's edits.
</Info>

### Quick Actions

The same **Editor**, **Trends**, **Clone**, and **Archive** actions available in the panel header are the primary entry points for the workflows below — use whichever one matches what you need to do with the selected agent.

## Clone Agents

Agent cloning now uses a guided **two-step assistant** instead of a single form, making it clear exactly what will be duplicated before anything is created.

<Steps>
  <Step title="Step 1 — Configure the Clone">
    Open **Clone** from the agent detail panel's header actions to fill in the base configuration:

    * **Agent Name**: defaults to `<Original Name> (Copy)`, fully editable
    * **Target Company**: shown only to admin users on the Meetzy domain with access to multiple companies; regular users automatically clone into their own company
    * **Phone Number**: select an available number from the same Twilio account as the original agent — the list loads automatically once a company is chosen
    * **Number Usage**: choose whether the assigned number is used for **Outbound & Inbound**, **Outbound only**, or **Inbound only**
    * **Agent Tools**: choose between **Duplicate tools** (independent copies, safe to edit separately) or **Reuse existing tools** (both agents share the same tools, and edits to one affect the other)
  </Step>

  <Step title="Step 2 — Review and Confirm">
    Clicking **Continue** moves to a summary screen listing the target company, the phone number and its usage mode, and the tools strategy in plain language before you confirm. Use **Back** to return and adjust any setting.
  </Step>

  <Step title="Cloning in Progress">
    Confirming triggers the clone request; the button switches to a disabled **"Cloning..."** state to prevent duplicate submissions. On success, the new agent appears in the list; on failure, an inline error message explains what went wrong so you can retry without losing your inputs.
  </Step>
</Steps>

<Warning>
  If the original agent's Twilio account has no available numbers, the modal shows a warning explaining that a free number from the **same account as the original agent** is required — release a number or purchase a new one before cloning.
</Warning>

<Note>
  The list of tools on the original agent is loaded and shown in the modal so you know exactly what will be duplicated or shared before confirming.
</Note>

## Archive Agents

<Info>
  Archiving is the recommended way to retire an agent you no longer need while keeping its configuration, call history, and analytics intact for future reference.
</Info>

Archiving deactivates an agent without deleting its data. The **Archive Agent** modal, available from the detail panel's header actions, guides you through the process:

<Steps>
  <Step title="Review the Confirmation">
    The modal explains that this will hide the agent from the agents list without deleting any of its data.
  </Step>

  <Step title="Choose Phone Number Handling">
    A dedicated checkbox lets you **"Release its phone number for other agents"**:

    * **Checked**: resets the Twilio configuration and frees the number so it can be assigned to another agent
    * **Unchecked (default)**: the number stays reserved with this agent
  </Step>

  <Step title="Confirm">
    Click **Archive** to proceed. The button shows an **"Archiving..."** loading state while the request completes. Any error is displayed inline in the modal so you can retry safely.
  </Step>
</Steps>

<Warning>
  Releasing the phone number is **irreversible** — once released, the number may be reassigned to a different agent or company. Leave the checkbox disabled if you plan to reactivate this agent later with the same number.
</Warning>

### Restoring an Archived Agent

Toggle **Archived agents** to **Shown** in the filters bar to locate an archived agent, select it, and restore it from its detail panel. If its phone number was released during archiving, assign a new one from the [Phone Numbers](/agents/phone-numbers) section before it can resume handling calls.

## Swap Phone Number

The **Swap Phone Number** modal lets you replace an agent's current number with a new one from the same Twilio account — useful when a number develops deliverability issues or needs to be freed up.

<Steps>
  <Step title="Open the Swap Modal">
    From the "Call" modal next to any phone number, or from the agent's Settings tab, click the swap icon next to the number you want to replace.
  </Step>

  <Step title="Review the Swap">
    The modal displays the **current number** (marked for release) and automatically loads the **next available number** from the same Twilio account to assign in its place.
  </Step>

  <Step title="Confirm">
    Click **Confirm Swap** to apply the change. The button enters a **"Swapping..."** state while the request is processed against both the database and Twilio. If no available numbers are found, an inline message tells you so, including the Twilio account label.
  </Step>
</Steps>

<Warning>
  A swap updates both your database and Twilio configuration immediately, and the old number is marked for release. This action cannot be undone from the modal.
</Warning>

## Custom Call Form Fields

Every agent that accepts outbound test/init calls has a configurable form. The **Configure Form Fields** editor lets you manage the custom fields shown in that form, beyond the built-in system fields.

<Info>
  System fields — **phone**, **voice**, and **language** — are always required and cannot be edited or removed from this editor.
</Info>

<Steps>
  <Step title="Open the Field Editor">
    Click **Edit Form** in the "Call" modal of any agent to open the **Configure Form Fields** modal.
  </Step>

  <Step title="Add or Edit Fields">
    For each custom field, configure:

    * **Field Label**: the text shown to users filling out the form
    * **Field Type**: Text, Email, Number, Phone, or Dropdown
    * **Parameter ID**: the identifier referenced in your agent's prompt as `{{id}}` — auto-generated from the label but fully editable
    * **Placeholder Text**: optional hint text shown inside the input
    * **Required field**: whether the field must be filled before a call can be started
  </Step>

  <Step title="Remove Fields">
    Click the trash icon on any field card to delete it (with a confirmation prompt).
  </Step>

  <Step title="Save">
    Click **Save** to persist the field configuration. Parameter IDs are validated for uniqueness and checked against system field names before saving; any conflict is reported inline so you can fix it without losing your changes.
  </Step>
</Steps>

<Tip>
  Custom Parameter IDs are automatically slugified from the field label (lowercase, no accents, underscores instead of spaces) unless you manually override them.
</Tip>

## Test Calling from the "Call" Modal

Each agent row includes a **Call** modal with everything needed to place a quick test call — inbound or outbound — without leaving the Agents page.

<Tabs>
  <Tab title="Inbound Testing">
    The modal lists the agent's assigned phone number(s) as clickable `tel:` links, plus a one-click **copy without country code** button for each number. If the agent only has verified caller IDs (no inbound number), a note clarifies it only supports outbound calls.
  </Tab>

  <Tab title="Outbound Testing">
    A dynamic form built from the agent's configured fields (including any custom fields added via the field editor), plus:

    * **Country prefix selector**: choose from a long list of country codes (with flags) next to the phone input — your last choice is remembered for next time via local storage
    * **Agent Version selector**: for Advanced agents only, choose a specific saved version to test against, or leave it blank to use the production version
    * **Edit Form** shortcut to open the custom fields editor directly from the modal
  </Tab>
</Tabs>

<Info>
  The **Call** button opens this flow as a standalone modal rather than an inline panel, so opening it never shifts the rows below it in the agent list — only one agent's Call modal can be open at a time.
</Info>

<Warning>
  Test calls consume billable minutes from your plan. Monitor usage in the billing dashboard.
</Warning>

## Permissions and Access Control

<Tabs>
  <Tab title="Admin Users">
    * Full access to clone, archive, restore, and swap phone numbers on any agent
    * Company filter available in both the agent list and the clone modal's target company selector
    * Cross-company agent management
    * Only role that sees the Twilio/ElevenLabs debugging links in the Calls tab
  </Tab>

  <Tab title="Advanced Editor Users">
    * Create and edit Advanced agents via the [Advanced Editor](/agents/advanced-editor)
    * Clone, archive, and manage phone numbers for their own agents
    * Access to the Settings tab for standard agents
  </Tab>

  <Tab title="Standard Users">
    * View assigned agents and their Overview/Calls tabs
    * Place test calls using the Call modal
    * Limited or no access to Archive, Clone, and Settings depending on role
  </Tab>
</Tabs>

## Next Steps

<CardGroup>
  <Card title="Phone Numbers" icon="phone" href="/agents/phone-numbers">
    Assign, release, and manage phone numbers for your agents.
  </Card>

  <Card title="Advanced Editor" icon="code" href="/agents/advanced-editor">
    Learn to use the Advanced Editor for full agent configuration.
  </Card>

  <Card title="Knowledge Base" icon="book-open" href="/agents/knowledge-base">
    Manage agent knowledge and information sources.
  </Card>

  <Card title="Analytics" icon="chart-line" href="/agents/analytics">
    Monitor agent performance and call metrics in depth.
  </Card>
</CardGroup>


# Phone Health
Source: https://docs.meetzy.es/agents/phone-health

Monitor your phone numbers health and avoid being marked as spam

## What is Phone Health?

**Phone Health** is a monitoring system that analyzes the reputation of phone numbers assigned to your agents. It helps you identify numbers being marked as spam and take preventive action.

<Info>
  Numbers marked as spam have significantly lower answer rates. Keeping numbers "healthy" is crucial for campaign success.
</Info>

<Info>
  The Phone Health report can now be accessed from two places in the new agents view: directly from the **agents list** via the status icon with the spam counter, and from the **Overview** tab of the new agent detail panel. Both entry points open the same report described below.
</Info>

## Understanding Number Types

Your agents can use two types of phone numbers for outbound calls:

### Twilio Numbers

* **Owned by Meetzy**: Numbers from our Twilio inventory
* **Badge**: Show "Twilio" badge in the health report
* **Management**: Can be swapped automatically for new numbers
* **Billing**: Included in your Meetzy subscription

### Caller IDs (Verified Numbers)

* **Client-owned**: Your own phone numbers verified through Twilio
* **Badge**: Show "Caller ID" badge in the health report
* **Management**: Can be disconnected but not automatically swapped
* **Purpose**: Use your existing business numbers for outbound calls

<Warning>
  Some agents may operate exclusively with Caller IDs and have no Twilio numbers assigned. These agents can only make outbound calls using verified client numbers.
</Warning>

## Phone Health Indicator

In the agents list, each agent with assigned numbers shows an indicator:

### Indicator Colors

| Color     | Meaning   | Affected Numbers     |
| --------- | --------- | -------------------- |
| 🟢 Green  | Healthy   | 0 numbers affected   |
| 🟡 Yellow | Attention | 1-2 numbers affected |
| 🟠 Orange | Warning   | 3-5 numbers affected |
| 🔴 Red    | Critical  | 6+ numbers affected  |

The number on the badge indicates how many numbers have spam issues.

## View Phone Health Report

You can now open the detailed report from two different places, whichever is more convenient in your workflow:

<Tabs>
  <Tab title="From the agents list">
    <Steps>
      <Step title="Locate the agent">
        Find the agent in the list
      </Step>

      <Step title="Click health indicator">
        Click the **signal/antenna icon** with the spam counter, shown next to the agent name
      </Step>

      <Step title="Review report">
        The **Phone Health Report** modal opens with detailed analysis
      </Step>
    </Steps>
  </Tab>

  <Tab title="From the agent detail panel">
    <Steps>
      <Step title="Select the agent">
        Click an agent in the list to open its detail panel
      </Step>

      <Step title="Go to Overview">
        The **Overview** tab is selected by default and shows a **Phone health** card
      </Step>

      <Step title="Open the report">
        Click the **Phone health** card to open the same **Phone Health Report** modal
      </Step>
    </Steps>
  </Tab>
</Tabs>

<Info>
  Both entry points show the exact same data — pick whichever is faster depending on whether you're scanning the whole list or already reviewing a single agent's details.
</Info>

## Report Content

### Summary Dashboard

The report shows three main metrics:

| Metric              | Description                                             |
| ------------------- | ------------------------------------------------------- |
| **Total Numbers**   | All numbers assigned to the agent (Twilio + Caller IDs) |
| **Healthy Numbers** | Numbers without spam issues                             |
| **Spam Numbers**    | Numbers with high spam ratio                            |

### Spam Threshold

```
Spam Threshold: Numbers with ≥50% spam ratio are considered problematic
```

A number is considered "spam" when more than 50% of its calls are marked or reported as spam.

### Detailed Analysis Table

Each number in the table shows:

| Column              | Description                       |
| ------------------- | --------------------------------- |
| **Phone Number**    | The phone number with type badge  |
| **Type Badge**      | "Twilio" or "Caller ID" indicator |
| **Status**          | Healthy, Warning, or Spam         |
| **Total Calls**     | Total calls made from this number |
| **Spam Calls**      | Calls marked as spam              |
| **Completed Calls** | Successfully completed calls      |
| **Spam Ratio**      | Spam percentage with visual bar   |
| **Actions**         | Available management options      |

### Number Status Types

| Status         | Spam Ratio    | Description               |
| -------------- | ------------- | ------------------------- |
| 🟢 **Healthy** | Less than 50% | Number is performing well |
| 🟡 **Warning** | 50% - 70%     | Take preventive action    |
| 🔴 **Spam**    | Above 70%     | Number should be replaced |

## Available Actions

### For Twilio Numbers

#### Swap Number

Replace a problematic Twilio number with a fresh one:

<Steps>
  <Step title="Identify number">
    Find a number with **Warning** or **Spam** status
  </Step>

  <Step title="Click Swap">
    Click the **Swap** button next to the number
  </Step>

  <Step title="Confirm replacement">
    The system automatically finds and assigns a new number
  </Step>

  <Step title="Complete swap">
    The old number is marked as "RETURN" and unlinked
  </Step>
</Steps>

<Warning>
  The old number will no longer be used by this agent after swapping.
</Warning>

#### Disconnect Number

Remove a Twilio number without replacing it:

<Steps>
  <Step title="Click Disconnect">
    Click the **Disconnect** button next to the number
  </Step>

  <Step title="Confirm action">
    Confirm the disconnection
  </Step>

  <Step title="Number released">
    The number becomes available for other agents
  </Step>
</Steps>

### For Caller IDs (Verified Numbers)

#### Disconnect Only

Caller IDs can only be disconnected (not swapped):

<Steps>
  <Step title="Click Disconnect">
    Click the **Disconnect** button next to the Caller ID
  </Step>

  <Step title="Confirm removal">
    Confirm the disconnection
  </Step>

  <Step title="Manual replacement">
    To replace, manually verify a new Caller ID in agent settings
  </Step>
</Steps>

<Info>
  Caller IDs are client-owned numbers and cannot be automatically replaced. You must verify new numbers manually if needed.
</Info>

## Best Practices

### Preventing Spam

<Steps>
  <Step title="Monitor regularly">
    Check Phone Health at least once a week
  </Step>

  <Step title="Act early">
    Swap Twilio numbers when they reach "Warning" (50%), don't wait for "Spam" (70%)
  </Step>

  <Step title="Distribute load">
    Use multiple numbers for high-volume campaigns
  </Step>

  <Step title="Respect schedules">
    Avoid calling at inappropriate times that generate more spam reports
  </Step>

  <Step title="Manage Caller IDs">
    Monitor verified numbers and disconnect problematic ones promptly
  </Step>
</Steps>

### Warning Signs

Watch for these indicators:

* Declining answer rates across campaigns
* Calls ending very quickly (under 3 seconds)
* Sudden increase in spam ratio
* Customer complaints about blocked numbers
* Poor performance from newly verified Caller IDs

## Mixed Number Strategy

For optimal performance, consider using both number types:

<Tabs>
  <Tab title="Twilio Numbers">
    **Best for:**

    * High-volume calling
    * Testing campaigns
    * Automated number rotation
    * Quick replacement when marked as spam
  </Tab>

  <Tab title="Caller IDs">
    **Best for:**

    * Brand recognition
    * Important client calls
    * Local presence
    * Regulatory compliance
  </Tab>
</Tabs>

## How Spam Ratio is Calculated

```
Spam Ratio = (Spam Calls / Total Calls) × 100
```

Calls are classified as "spam" based on:

* Direct user reports to carriers
* Public spam databases
* Extremely short duration (\< 3 seconds)
* Systematic rejection patterns
* Carrier-level blocking

## Top Performer Number

In the report, the best-performing number is marked with:

```
🏆 Top Performer
```

This number has the best completion history and should be prioritized for important calls.

## FAQ

<AccordionGroup>
  <Accordion title="Can I swap a Caller ID like a Twilio number?">
    No, Caller IDs are client-owned and cannot be automatically swapped. You can only disconnect them and manually verify new ones if needed.
  </Accordion>

  <Accordion title="Why does my agent only have Caller IDs?">
    Some agents are configured to use only verified client numbers for outbound calls. This is common for compliance or branding requirements.
  </Accordion>

  <Accordion title="How often is Phone Health updated?">
    Data is updated every hour. The indicator in the agents list reflects the latest available analysis.
  </Accordion>

  <Accordion title="Can I recover a Twilio number after swapping?">
    Not automatically. Once swapped, the number is unlinked. Contact support if you need to recover it.
  </Accordion>

  <Accordion title="What happens to active calls during actions?">
    Active calls are not interrupted. Changes apply only to new calls.
  </Accordion>

  <Accordion title="Why does my new Caller ID already have spam history?">
    Verified numbers may inherit reputation from previous usage. Monitor new Caller IDs closely and disconnect if problematic.
  </Accordion>

  <Accordion title="Why do I see the Phone Health report in two places now?">
    The report is now reachable both from the status icon in the agents list and from the Overview tab of the agent detail panel, so you can check it whether you're browsing the full list or already inspecting a specific agent. It's the same report in both cases.
  </Accordion>
</AccordionGroup>

## Next Steps

<CardGroup>
  <Card title="Configure Numbers" icon="phone" href="/agents/phone-numbers">
    Learn to assign and manage phone numbers and Caller IDs.
  </Card>

  <Card title="Manage Agents" icon="list" href="/agents/manage-agents">
    View and organize your agents.
  </Card>
</CardGroup>


# Configure Phone Numbers
Source: https://docs.meetzy.es/agents/phone-numbers

Learn how to assign, manage, and swap phone numbers for your Meetzy agents

## Phone Numbers in Meetzy

Each Meetzy agent needs at least one phone number to:

* **Receive inbound calls**
* **Make outbound calls**
* **Identify itself** (caller ID)

<Info>
  Phone numbers are exclusive resources. Each number can only be assigned to one agent at a time.
</Info>

## Phone Number Types

Meetzy supports two types of phone numbers for your agents:

### Twilio Phone Numbers

Standard phone numbers managed through your Twilio account:

* Automatically provisioned and configured
* Support both inbound and outbound calls
* Can be swapped and reassigned automatically
* Managed through Meetzy's interface

### Verified Caller IDs

Client-owned phone numbers verified for outbound calling:

* Numbers you already own and want to use as caller ID
* Must be verified through Twilio's verification process
* Appear with a special verification badge in the interface
* Cannot be swapped automatically (manual process required)

<Warning>
  Caller IDs are for outbound calling only and cannot be automatically rotated like Twilio numbers. They require manual management for spam protection.
</Warning>

## Visual Identification

In the Meetzy interface, you can distinguish between phone number types:

| Type                   | Visual Indicator                     | Description                  |
| ---------------------- | ------------------------------------ | ---------------------------- |
| **Twilio Number**      | Standard format with swap controls   | Fully managed by Meetzy      |
| **Verified Caller ID** | 🔒 Lock icon with verification badge | Client-owned verified number |

<Info>
  Verified Caller IDs show verification status badges (pending, verified, disabled) to help you track their state.
</Info>

## Multi-Subaccount Phone Number Management

Meetzy now supports managing phone numbers across multiple Twilio subaccounts with enhanced assignment states and improved management features.

### Assignment States

Phone numbers can have three distinct assignment states:

| State          | Description                                 | Visual Indicator |
| -------------- | ------------------------------------------- | ---------------- |
| **Assigned**   | Fully assigned to an agent and active       | 🟢 Green badge   |
| **Partial**    | Partially assigned or pending configuration | 🟡 Yellow badge  |
| **Unassigned** | Available for assignment                    | ⚪ Gray badge     |

### Subaccount Credentials Management

Each Twilio subaccount can have its own set of credentials configured independently:

<Tabs>
  <Tab title="Main Account">
    Your primary Twilio account with full management capabilities
  </Tab>

  <Tab title="Configured Subaccounts">
    Subaccounts with proper credentials - full management available
  </Tab>

  <Tab title="Unconfigured Subaccounts">
    Subaccounts without credentials - view-only access
  </Tab>
</Tabs>

<Warning>
  To fully manage phone numbers in a subaccount, you need to configure the subaccount credentials in your settings.
</Warning>

## Twilio Subaccount Navigation

The phone numbers interface includes a tabbed navigation system to manage numbers across different Twilio subaccounts:

### How to Navigate Between Subaccounts

<Steps>
  <Step title="View Available Subaccounts">
    Look for the tab interface at the top of the phone numbers page.
  </Step>

  <Step title="Select Subaccount">
    Click on any tab to switch to that subaccount's phone numbers.
  </Step>

  <Step title="Wait for Loading">
    The system will load numbers for the selected subaccount automatically.
  </Step>

  <Step title="Check Credential Status">
    Look for the credential status badge on each subaccount tab.
  </Step>
</Steps>

<Info>
  Phone numbers are cached per subaccount to improve performance when switching between tabs.
</Info>

## Enhanced Phone Number Filtering

The interface provides several filtering options to help you manage large numbers of phone numbers:

### Filter Options

| Filter         | Description                                  | Use Case                  |
| -------------- | -------------------------------------------- | ------------------------- |
| **All**        | Shows all phone numbers regardless of status | Complete overview         |
| **Assigned**   | Shows only numbers assigned to agents        | Review active assignments |
| **Unassigned** | Shows only unassigned numbers                | Find available numbers    |

### Visual Status Indicators

Each phone number displays color-coded status indicators based on the new assignment states:

* 🟢 **Green**: Fully assigned and active
* 🟡 **Yellow**: Partially assigned or pending
* ⚪ **Gray**: Unassigned and available

## Direct Agent Assignment System

The new assignment system allows you to directly assign phone numbers to agents with improved feedback:

<Steps>
  <Step title="Locate Available Number">
    Find an unassigned number in the list.
  </Step>

  <Step title="Click Assign">
    Click the "Assign" button next to the phone number.
  </Step>

  <Step title="Select Agent">
    Choose an agent from the dropdown list of available agents.
  </Step>

  <Step title="Confirm Assignment">
    Click confirm to complete the assignment with automatic state updates.
  </Step>
</Steps>

### Assignment Process

When you assign a number:

1. The number's status changes to **assigned**
2. The agent's configuration is updated
3. The cache is refreshed for real-time updates
4. Assignment status is reflected across all relevant interfaces

<Warning>
  Once assigned directly, the number will become unavailable for other agents until reassigned, reset, or returned.
</Warning>

## Advanced Number Management Operations

### Reset Numbers

The reset function allows you to clear a number's configuration while keeping it in your account:

<Steps>
  <Step title="Select Number to Reset">
    Find the phone number you want to reset.
  </Step>

  <Step title="Click Reset">
    Use the reset option in the number's actions menu.
  </Step>

  <Step title="Confirm Reset">
    Confirm the reset operation - this will clear webhooks and assignments.
  </Step>

  <Step title="Verify Status">
    The number will return to **unassigned** status and be available for new assignments.
  </Step>
</Steps>

### Return Numbers

The return function removes numbers from your account completely:

<Steps>
  <Step title="Select Number to Return">
    Find the phone number you want to return to Twilio's pool.
  </Step>

  <Step title="Click Return">
    Use the return option in the number's actions menu.
  </Step>

  <Step title="Confirm Return">
    Confirm the return operation - this action cannot be undone.
  </Step>

  <Step title="Number Released">
    The number is returned to Twilio and removed from your account.
  </Step>
</Steps>

<Warning>
  Returned numbers cannot be recovered. They may be assigned to other customers.
</Warning>

## Assign a Number When Creating an Agent

When creating a new agent in the Advanced Editor:

1. Go to the **Settings** section (Cog icon)
2. Find the **Phone Number** field
3. Select an available number from the dropdown

### If No Numbers Are Available

If the dropdown is empty, you'll see a message:

```
No available phone numbers for the selected account.
All numbers are assigned or configured.
```

**Options:**

* Release a number from another agent
* Reset unused numbers
* Request new numbers via the "Request more numbers" link

## View Assigned Numbers

Once assigned, the number appears in read-only mode:

```
Phone Number: +1 555 123 4567 (Sales Team) 🔒
Phone number assigned to this agent
```

<Warning>
  The phone number cannot be changed directly once assigned. To change it, use the Swap function from Phone Health.
</Warning>

## Caller ID Management

Verified Caller IDs provide additional flexibility for outbound calling with your own phone numbers.

### Adding a Caller ID

<Steps>
  <Step title="Navigate to Agent Settings">
    Open the agent configuration in the Advanced Editor.
  </Step>

  <Step title="Add Caller ID">
    In the Caller IDs section, enter your phone number.
  </Step>

  <Step title="Start Verification">
    Twilio will initiate the verification process.
  </Step>

  <Step title="Complete Verification">
    Follow the verification steps provided by Twilio.
  </Step>
</Steps>

### Caller ID Limitations

<Warning>
  Verified Caller IDs have specific limitations compared to Twilio numbers:

  * Cannot be automatically swapped for spam protection
  * Require manual management if flagged
  * Only support outbound calls as caller ID
  * Must be verified through Twilio's process
</Warning>

### Managing Caller ID Status

Caller IDs can have different verification states:

| Status       | Description                    | Actions Available             |
| ------------ | ------------------------------ | ----------------------------- |
| **Pending**  | Verification in progress       | Complete verification process |
| **Verified** | Ready for use                  | Use for outbound calls        |
| **Disabled** | Verification failed or revoked | Re-verify or remove           |

## Multiple Numbers per Agent

An agent can have multiple assigned numbers. This is useful for:

* **Volume distribution**: Reduce spam risk
* **Regional segmentation**: Different numbers for different areas
* **Redundancy**: Backup if a number has issues

### How Multiple Numbers Are Used

| Call Type    | Behavior                                               |
| ------------ | ------------------------------------------------------ |
| **Inbound**  | Customer can call any of the numbers                   |
| **Outbound** | System automatically selects the best available number |

The number with the best **Phone Health** has priority for outbound calls.

## Swap Numbers

You can swap a problematic number for a new one from the Phone Health report:

<Steps>
  <Step title="Open Phone Health">
    Click the health indicator on the agent.
  </Step>

  <Step title="Identify the number">
    Find the number with Warning or Spam status.
  </Step>

  <Step title="Start Swap">
    Click the **Swap** button next to the number.
  </Step>

  <Step title="Confirm">
    The system shows the proposed new number. Confirm the swap.
  </Step>
</Steps>

### What Happens During a Swap

1. **Old number** is unlinked and marked for return
2. **New number** is assigned with **assigned** status
3. Assignment states are updated in real-time

<Info>
  Only Twilio numbers can be swapped automatically. Verified Caller IDs must be managed manually.
</Info>

## Disconnect a Number

To remove a number without replacing it:

1. Open the **Phone Health** report
2. Click **Disconnect** next to the number
3. Confirm the disconnection

<Warning>
  If you disconnect the agent's only number, it won't be able to make or receive calls until you assign another.
</Warning>

## Request New Numbers

To request additional numbers:

1. In the Advanced Editor, Settings section
2. Click **"Request more numbers"**
3. You'll be redirected to the request portal

New numbers are typically available within 24-48 hours.

## Supported Country Prefixes

Meetzy supports phone numbers in multiple countries:

| Country        | Prefix | Availability |
| -------------- | ------ | ------------ |
| United States  | +1     | Available    |
| Spain          | +34    | Available    |
| Mexico         | +52    | Available    |
| Argentina      | +54    | Available    |
| Colombia       | +57    | Available    |
| United Kingdom | +44    | Available    |

<Info>
  Use local numbers from your customers' country to increase answer rates. Customers are more likely to answer calls from local numbers.
</Info>

## Best Practices

### Number Management Workflow

For efficient number management:

<Steps>
  <Step title="Regular Monitoring">
    Check assignment states regularly across all subaccounts.
  </Step>

  <Step title="Credential Management">
    Ensure all subaccounts have proper credentials configured.
  </Step>

  <Step title="Assignment Tracking">
    Use the new assignment states to track number utilization.
  </Step>

  <Step title="Cleanup Operations">
    Use reset and return functions to manage unused numbers.
  </Step>
</Steps>

### Number Rotation

For high-volume campaigns:

1. Assign multiple numbers to the agent
2. The system rotates automatically based on assignment status
3. Monitor Phone Health regularly
4. Swap numbers before they reach "Spam" status

### Caller ID Strategy

For verified Caller IDs:

1. Use for brand recognition and customer trust
2. Monitor spam reports manually
3. Have backup Twilio numbers ready
4. Keep verification status current

### Subaccount Organization

* Use the main account for primary business numbers
* Organize specialized campaigns in separate subaccounts
* Apply consistent naming conventions across subaccounts
* Configure credentials for full management capabilities
* Regularly review and clean up unused numbers using reset/return functions

## FAQ

<AccordionGroup>
  <Accordion title="What's the difference between Twilio numbers and Caller IDs?">
    **Twilio numbers** are fully managed by Meetzy and can be automatically swapped. **Verified Caller IDs** are your own phone numbers that require verification and manual management.
  </Accordion>

  <Accordion title="Can Caller IDs be swapped automatically?">
    No, verified Caller IDs cannot be swapped automatically like Twilio numbers. They require manual management if they become flagged for spam.
  </Accordion>

  <Accordion title="How do I identify Caller IDs in the interface?">
    Caller IDs appear with a 🔒 lock icon and verification status badge, while Twilio numbers show standard management controls.
  </Accordion>

  <Accordion title="What's the difference between reset and return?">
    **Reset** clears a number's configuration but keeps it in your account as unassigned. **Return** completely removes the number from your account and returns it to Twilio's pool.
  </Accordion>

  <Accordion title="Can I recover a returned number?">
    No, returned numbers cannot be recovered. They become available to other Twilio customers.
  </Accordion>

  <Accordion title="What happens if I don't configure subaccount credentials?">
    You can view phone numbers in unconfigured subaccounts but cannot perform management operations like assign, reset, or return.
  </Accordion>

  <Accordion title="How do assignment states help with management?">
    Assignment states provide clear visibility into number utilization and help identify numbers that need attention or can be reset/returned.
  </Accordion>

  <Accordion title="Are numbers exclusive or shared?">
    Each number is exclusive to your account. It's not shared between different Meetzy customers.
  </Accordion>

  <Accordion title="Can I have numbers from different countries on one agent?">
    Yes, you can assign numbers from multiple countries to the same agent.
  </Accordion>

  <Accordion title="Do subaccounts share phone numbers?">
    No, each subaccount maintains its own pool of phone numbers that cannot be shared with other subaccounts.
  </Accordion>

  <Accordion title="How do I move a number between subaccounts?">
    Numbers cannot be moved between subaccounts. You would need to return the number and request a new one in the target subaccount.
  </Accordion>
</AccordionGroup>

## Next Steps

<CardGroup>
  <Card title="Phone Health" icon="heart-pulse" href="/agents/phone-health">
    Monitor the health and reputation of your numbers.
  </Card>

  <Card title="Make Test Calls" icon="phone-arrow-up-right" href="/agents/manage-agents#make-test-calls">
    Learn to test your agent with real calls.
  </Card>
</CardGroup>


# Create a campaign
Source: https://docs.meetzy.es/api-reference/endpoint/campaign

POST /campaign
Create outbound call campaigns from a contact list (immediate or scheduled)

## Overview

This endpoint creates outbound call campaigns using a playbook (assistant) and a list of contacts. Campaigns can run **immediately** or be **scheduled** for a specific date and time. Contacts are created or updated in the system and calls are launched according to the campaign configuration.

***

## 1. Summary

* **Endpoint:** `POST /campaign`
* **Authentication:** Bearer token (same as for the rest of the API).
* **Usage:** Send a playbook ID, a list of contacts (with at least `phone`), and optionally a launch date and recall policy. The system creates or updates contacts and launches the calls.

***

## 2. Authentication

Include your token in the request headers:

```http theme={null}
Authorization: Bearer YOUR_API_TOKEN
Content-Type: application/json
```

The same token is used for **POST /call** and **GET /call/:id**.

***

## 3. Request body

### 3.1 Required fields

| Field         | Type   | Description                                                                       |
| ------------- | ------ | --------------------------------------------------------------------------------- |
| `playbook_id` | string | UUID of the playbook (assistant) used for the calls. Must belong to your account. |
| `contacts`    | array  | List of contacts to call. Each item must have at least the `phone` field.         |

### 3.2 Optional fields (campaign level)

| Field       | Type            | Description                                                                                                                                         |
| ----------- | --------------- | --------------------------------------------------------------------------------------------------------------------------------------------------- |
| `launch_at` | string          | Launch date/time in ISO format (e.g. `"2025-03-01T10:00:00"`). If provided, the campaign is scheduled for that time; otherwise it runs immediately. |
| `recall`    | string or array | Retry policy in minutes. See section 4.                                                                                                             |

### 3.3 Optional fields per contact

Each item in `contacts` may include, in addition to `phone`, any of these fields (for call personalization or CRM):

| Field           | Type   | Description                     |
| --------------- | ------ | ------------------------------- |
| `first_name`    | string | First name                      |
| `last_name`     | string | Last name                       |
| `email`         | string | Email                           |
| `company`       | string | Company                         |
| `job_title`     | string | Job title                       |
| `address`       | string | Address                         |
| `city`          | string | City                            |
| `state`         | string | State / province                |
| `country`       | string | Country                         |
| `postal_code`   | string | Postal code                     |
| `timezone`      | string | Timezone (e.g. `Europe/Madrid`) |
| `utm_source`    | string | UTM source                      |
| `utm_medium`    | string | UTM medium                      |
| `utm_campaign`  | string | UTM campaign                    |
| `utm_term`      | string | UTM term                        |
| `utm_content`   | string | UTM content                     |
| `custom_fields` | object | Custom key-value fields         |

***

## 4. Using the `recall` field

The `recall` field defines **retry delays in minutes** from the time of the call (or from launch time for scheduled campaigns). It is stored per call and used to automatically reschedule calls (e.g. if the contact does not answer).

**String with array of minutes:**

`"recall": "[30,120]"` — retry at 30 and 120 minutes.

Example in the body:

```json theme={null}
{
  "playbook_id": "550e8400-e29b-41d4-a716-446655440000",
  "contacts": [ { "phone": "+34600123456", "first_name": "Juan" } ],
  "recall": "[30,120]"
}
```

***

## 5. Request examples

### 5.1 Immediate campaign (minimal)

Only playbook and list of phones:

```json theme={null}
{
  "playbook_id": "550e8400-e29b-41d4-a716-446655440000",
  "contacts": [
    { "phone": "+34600123456" },
    { "phone": "+34600987654" }
  ]
}
```

### 5.2 Immediate campaign with contact data and recall

```json theme={null}
{
  "playbook_id": "550e8400-e29b-41d4-a716-446655440000",
  "contacts": [
    {
      "phone": "+34600123456",
      "first_name": "Juan",
      "last_name": "García",
      "email": "juan@example.com",
      "company": "Acme SL"
    },
    {
      "phone": "+34600987654",
      "first_name": "María",
      "email": "maria@example.com"
    }
  ],
  "recall": "[30,120]"
}
```

### 5.3 Immediate campaign with custom fields

```json theme={null}
{
  "playbook_id": "550e8400-e29b-41d4-a716-446655440000",
  "contacts": [
    {
      "phone": "+34600123456",
      "first_name": "Juan",
      "last_name": "García",
      "custom_fields": {
        "lead_id": "ext-12345",
        "producto_interes": "Premium",
        "origen": "web"
      }
    }
  ],
  "recall": [30, 60, 120]
}
```

### 5.4 Scheduled campaign

Use `launch_at` to schedule the launch:

```json theme={null}
{
  "playbook_id": "550e8400-e29b-41d4-a716-446655440000",
  "contacts": [
    { "phone": "+34600123456", "first_name": "Juan" },
    { "phone": "+34600987654", "first_name": "María" }
  ],
  "launch_at": "2025-03-01T10:00:00",
  "recall": "[30,120]"
}
```

Calls are created at the given date/time (with a small stagger between contacts). `recall` works the same as for immediate campaigns.

***

## 6. Server responses

### 6.1 Campaign accepted (immediate)

* **Status:** `202 Accepted`
* **Example body:**

```json theme={null}
{
  "success": true,
  "message": "Campaign started successfully",
  "campaign_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "scheduled": false,
  "total_contacts": 2
}
```

### 6.2 Scheduled campaign accepted

* **Status:** `202 Accepted`
* **Example body:**

```json theme={null}
{
  "success": true,
  "message": "Scheduled 2 calls for 01/03/2025 10:00",
  "campaign_id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "scheduled": true,
  "launch_at": "2025-03-01T10:00:00.000Z",
  "total_contacts": 2,
  "scheduled_calls": 2,
  "error_calls": 0
}
```

***

## 7. Error codes

| Code    | Meaning                                                                                      |
| ------- | -------------------------------------------------------------------------------------------- |
| **400** | Invalid data: missing `playbook_id`, empty `contacts`, or a contact without a valid `phone`. |
| **401** | Not authenticated: missing or invalid token.                                                 |
| **403** | Playbook not found or not owned by your account.                                             |
| **500** | Internal server error.                                                                       |

Example 400 response:

```json theme={null}
{
  "error": "Each contact must have a non-empty phone field"
}
```

***

## 8. Quick reference

1. **Authentication:** `Authorization: Bearer YOUR_TOKEN` header.
2. **Required:** `playbook_id` (UUID) and `contacts` (array of objects with at least `phone`).
3. **Optional:** `launch_at` to schedule; `recall` (e.g. `"[30,120]"`) for retries in minutes.
4. **Response:** `202` with `campaign_id`.


# Init a call
Source: https://docs.meetzy.es/api-reference/endpoint/create

POST /call
API that start a call using a pre-trained playbook.

## Overview

This endpoint initiates an outbound call using a pre-trained playbook. The call will be made using your configured phone number and the AI will follow the playbook's instructions.

## Version Control

By default, this endpoint uses the **production version** of your playbook. This ensures that all API calls use the same configuration you've designated as production-ready in the dashboard.

<Tip>
  You can optionally specify a `version_id` parameter to use a specific version of your playbook. This is useful for testing new versions before marking them as production. See [Agent Versioning](/api-reference/endpoint/versioning) for more details.
</Tip>

### Default Behavior

When you call this endpoint without a `version_id`:

1. The system looks for the playbook's **production version**
2. If found, it uses that version's configuration (prompt, voice, settings, etc.)
3. If no versions exist, it falls back to the current playbook configuration

### Using a Specific Version

```json theme={null}
{
  "phone": "+34600000000",
  "playbook_id": "your-playbook-id",
  "version_id": "optional-version-uuid"
}
```

## Response

A successful response includes the call ID which you can use to track the call status:

```json theme={null}
{
  "success": true,
  "id": "call-uuid",
  "message": "Call processing started"
}
```


# Get a call
Source: https://docs.meetzy.es/api-reference/endpoint/get

GET /call/{callId}
Returns data about a specific call.



# List Calls
Source: https://docs.meetzy.es/api-reference/endpoint/list

GET /calls
Retrieve a list of calls with filtering options

## Overview

This endpoint retrieves a list of calls associated with the authenticated user's assistants. You can filter calls by date range, status, and playbook.

## Authentication

This endpoint requires authentication using a Bearer token in the Authorization header.

```bash theme={null}
Authorization: Bearer YOUR_API_TOKEN
```

## Query Parameters

<ParamField type="string">
  Start date for filtering calls. Format: `YYYY-MM-DD` or `YYYY-MM-DD HH:mm:ss`

  If not provided, it will be calculated based on `daysFromNow` parameter.
</ParamField>

<ParamField type="string">
  End date for filtering calls. Format: `YYYY-MM-DD` or `YYYY-MM-DD HH:mm:ss`

  If not provided, it will default to the current date/time.
</ParamField>

<ParamField type="number">
  Number of days to look back from today. Only used if `startDate` and `endDate` are not provided.

  * Minimum: 1
  * Maximum: 90
  * Default: 15
</ParamField>

<ParamField type="string">
  Filter calls by status. Valid values:

  * `open` - Call is in progress or pending
  * `closed` - Call has been completed successfully
  * `rong-phone-forno-answer` - Wrong phone number or no answer
  * `wmat` - Waiting for manual action
  * `failed` - Call failed
  * `recall-scheduled` - A recall has been scheduled
  * `voicemail` - Call went to voicemail
</ParamField>

<ParamField type="string">
  Filter calls by a specific playbook (assistant) ID. Must be a valid UUID.
</ParamField>

## Response

<ResponseField name="calls" type="array">
  Array of call objects, ordered by `created_time` descending (most recent first).

  <Expandable title="Call Object">
    <ResponseField name="id" type="string">
      Unique identifier for the call (UUID)
    </ResponseField>

    <ResponseField name="created_time" type="string">
      When the call was created. Format: `YYYY-MM-DD HH:mm:ss` (UTC)
    </ResponseField>

    <ResponseField name="timestamp" type="number">
      Unix timestamp in milliseconds of when the call was created
    </ResponseField>

    <ResponseField name="assistant_id" type="string">
      ID of the assistant (playbook) that handled the call
    </ResponseField>

    <ResponseField name="from" type="string">
      Phone number that initiated the call. Empty if not available or invalid.
    </ResponseField>

    <ResponseField name="to" type="string">
      Phone number that received the call
    </ResponseField>

    <ResponseField name="status" type="string">
      Current status of the call. See status parameter for possible values.
    </ResponseField>

    <ResponseField name="duration" type="number">
      Duration of the call in seconds
    </ResponseField>

    <ResponseField name="recording" type="string">
      URL to the call recording (if available)
    </ResponseField>

    <ResponseField name="transcript" type="string">
      HTML-formatted transcript of the call. Includes speaker labels formatted as `<b>[Speaker]</b>`.
    </ResponseField>

    <ResponseField name="success" type="boolean">
      Whether the call was successful
    </ResponseField>

    <ResponseField name="summary" type="string">
      AI-generated summary of the call
    </ResponseField>

    <ResponseField name="output" type="object">
      Structured output data from the call. Format depends on the playbook configuration.
    </ResponseField>

    <ResponseField name="data" type="object">
      Additional call data and parameters. Sensitive fields like voice\_id, voice\_model, text\_model, and assistant\_id are removed for security.

      <Expandable title="Data Object">
        <ResponseField name="params" type="object">
          Custom parameters passed to the call
        </ResponseField>
      </Expandable>
    </ResponseField>
  </Expandable>
</ResponseField>

## Example Requests

### Get calls from the last 30 days

```bash theme={null}
GET https://api.meetzy.io/calls?daysFromNow=30
```

### Get calls within a specific date range

```bash theme={null}
GET https://api.meetzy.io/calls?startDate=2024-01-01&endDate=2024-01-31
```

### Get only closed calls

```bash theme={null}
GET https://api.meetzy.io/calls?status=closed
```

### Get calls for a specific playbook

```bash theme={null}
GET https://api.meetzy.io/calls?playbook_id=550e8400-e29b-41d4-a716-446655440000
```

### Combine filters

```bash theme={null}
GET https://api.meetzy.io/calls?daysFromNow=7&status=closed&playbook_id=550e8400-e29b-41d4-a716-446655440000
```

## Example Response

```json theme={null}
[
  {
    "id": "123e4567-e89b-12d3-a456-426614174000",
    "created_time": "2024-11-12 10:30:00",
    "timestamp": 1731407400000,
    "assistant_id": "550e8400-e29b-41d4-a716-446655440000",
    "from": "+34612345678",
    "to": "+34987654321",
    "status": "closed",
    "duration": 180,
    "recording": "https://recordings.meetzy.io/call-123.mp3",
    "transcript": "<b>[Agent]</b> Hello, this is Sarah from Meetzy...<br><br><b>[Customer]</b> Hi Sarah...",
    "success": true,
    "summary": "Customer was interested in the premium plan...",
    "output": {
      "appointment_scheduled": true,
      "date": "2024-11-15",
      "time": "14:00"
    },
    "data": {
      "params": {
        "campaign_id": "summer-2024",
        "lead_source": "website"
      }
    }
  }
]
```

## Error Responses

<ResponseField name="401 Unauthorized" type="object">
  Authentication failed or token is invalid

  ```json theme={null}
  {
    "error": "Unauthorized"
  }
  ```
</ResponseField>

<ResponseField name="403 Forbidden" type="object">
  User does not have permission to access this resource

  ```json theme={null}
  "Unauthorized"
  ```
</ResponseField>

<ResponseField name="500 Internal Server Error" type="object">
  An error occurred on the server

  ```json theme={null}
  "Internal Server Error"
  ```
</ResponseField>

## Notes

* All dates are handled in Europe/Madrid timezone and converted to UTC in the response
* The `from` field is automatically sanitized to remove invalid entries like "phoneNumber.phoneNumber"
* Transcripts are formatted with HTML tags for better readability in UI applications
* The endpoint automatically filters calls to only show those belonging to assistants owned by the authenticated user
* Maximum date range when using `daysFromNow` is 90 days


# Agent Versioning
Source: https://docs.meetzy.es/api-reference/endpoint/versioning

Understand how version control works for your playbooks and how to use specific versions in API calls.

## Overview

Meetzy supports version control for playbooks, allowing you to:

* Keep a history of all changes made to your playbook
* Designate a specific version as "production"
* Test specific versions before deploying to production
* Use historical versions for specific API calls

## How Versioning Works

### Production Version

Every playbook has one version marked as **production**. This is the version that will be used for:

* **Inbound calls**: All incoming calls to your Meetzy number
* **Outbound calls via API**: When you don't specify a `version_id`
* **Queued/scheduled calls**: Uses the version that was marked as production when the call was scheduled

### Version Selection Priority

When a call is initiated, the system follows this priority:

1. **Specific version requested**: If you provide `version_id` in the API call, that exact version is used
2. **Production version**: If no `version_id` is provided, the production version is used
3. **Current configuration (fallback)**: If no versions exist for the playbook, the current playbook configuration is used

## Using Versions in API Calls

### Default Behavior (Recommended)

Simply call the API without specifying a version. The production version will be used automatically:

```bash theme={null}
curl -X POST https://api.meetzy.io/call \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "phone": "+34600000000",
    "playbook_id": "your-playbook-id",
    "firstname": "John",
    "lastname": "Doe"
  }'
```

### Using a Specific Version

To use a specific version (e.g., for testing), include the `version_id` parameter:

```bash theme={null}
curl -X POST https://api.meetzy.io/call \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "phone": "+34600000000",
    "playbook_id": "your-playbook-id",
    "version_id": "abc123-version-uuid",
    "firstname": "John",
    "lastname": "Doe"
  }'
```

<Note>
  The `version_id` must belong to the specified `playbook_id`. If the version doesn't exist or belongs to a different playbook, the API will fall back to the production version.
</Note>

## Response

When you retrieve call details, the response includes the `version_id` that was used:

```json theme={null}
{
  "id": "call-uuid",
  "playbook_id": "your-playbook-id",
  "version_id": "abc123-version-uuid",
  "status": "completed",
  "duration": 120,
  "...": "..."
}
```

If `version_id` is empty or null, it means either:

* No version control was active for this playbook at the time
* The call used the legacy system (before version control was implemented)

## Managing Versions in the Dashboard

You can manage playbook versions in the Advanced Editor:

1. **View History**: See all committed versions with their timestamps and commit messages
2. **Mark as Production**: Designate any version as the production version
3. **Restore Version**: Revert to a previous version's configuration
4. **Compare Changes**: View differences between versions

<Warning>
  Changing the production version takes effect immediately for all new inbound calls and API calls that don't specify a version.
</Warning>

## Best Practices

1. **Always commit changes before going live**: Use the commit system to create a snapshot before major changes
2. **Test before marking as production**: Use the `version_id` parameter to test new versions with specific calls before making them production
3. **Use descriptive commit messages**: This helps track what changed in each version
4. **Keep production stable**: Only mark thoroughly tested versions as production

## Use Cases

### A/B Testing

Test a new greeting or prompt with specific leads before rolling out to everyone:

```bash theme={null}
# Test call with new version
curl -X POST https://api.meetzy.io/call \
  -d '{
    "phone": "+34600000000",
    "playbook_id": "your-playbook-id",
    "version_id": "new-greeting-version-id"
  }'
```

### Rollback

If you notice issues with a new production version, you can:

1. Go to the Advanced Editor → History tab
2. Find a known-good version
3. Click "Mark as Prod" to instantly roll back

### Scheduled Campaigns

When scheduling bulk calls, you can specify a version to ensure consistency even if you update the playbook later:

```bash theme={null}
# Schedule calls with a locked version
curl -X POST https://api.meetzy.io/call \
  -d '{
    "phone": "+34600000000",
    "playbook_id": "your-playbook-id",
    "version_id": "campaign-v1-version-id"
  }'
```

## FAQ

<AccordionGroup>
  <Accordion title="What happens if I don't have any versions?">
    The system falls back to using the current playbook configuration. This ensures backwards compatibility with playbooks created before version control was implemented.
  </Accordion>

  <Accordion title="Can I delete versions?">
    Currently, versions cannot be deleted to maintain audit trail integrity. You can always mark a different version as production.
  </Accordion>

  <Accordion title="Is there a limit to the number of versions?">
    There is no hard limit on the number of versions. However, we recommend keeping meaningful commits rather than committing every small change.
  </Accordion>

  <Accordion title="Do inbound calls use version control?">
    Yes! Inbound calls always use the production version. This ensures your customers always experience the configuration you've designated as production-ready.
  </Accordion>
</AccordionGroup>


# Introduction
Source: https://docs.meetzy.es/api-reference/introduction

Explore our API Reference to integrate Meetzy with your systems.

## Welcome to the Meetzy API

The Meetzy API allows you to programmatically trigger calls, retrieve call data, and integrate AI-powered voice agents with your existing systems.

## Base URL

All API requests are made to:

```
https://api.meetzy.io
```

## Authentication

All API requests require authentication using a Bearer token.

### Getting Your API Token

1. Log in to your [Meetzy Dashboard](https://dashboard.meetzy.io)
2. Navigate to **Account** → **API Tokens**
3. Copy your API token

### Using Your Token

Include the token in the `Authorization` header:

```bash theme={null}
curl -X GET https://api.meetzy.io/calls \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -H "Content-Type: application/json"
```

<Warning>
  Keep your API token secret. Never expose it in client-side code or public repositories.
</Warning>

## Request Format

All requests should include the `Content-Type: application/json` header when sending data.

```json theme={null}
{
  "phone": "+1234567890",
  "playbook_id": "pb_abc123",
  "params": {
    "customer_name": "John Doe"
  }
}
```

## Response Format

All responses are returned in JSON format:

```json theme={null}
{
  "success": true,
  "data": {
    "call_id": "call_xyz789",
    "status": "initiated"
  }
}
```

### Error Responses

Errors include an error code and message:

```json theme={null}
{
  "success": false,
  "error": {
    "code": "invalid_phone",
    "message": "The provided phone number is invalid"
  }
}
```

## HTTP Status Codes

| Code | Description                             |
| ---- | --------------------------------------- |
| 200  | Success                                 |
| 201  | Created                                 |
| 400  | Bad Request - Invalid parameters        |
| 401  | Unauthorized - Invalid or missing token |
| 403  | Forbidden - Insufficient permissions    |
| 404  | Not Found - Resource doesn't exist      |
| 429  | Too Many Requests - Rate limited        |
| 500  | Internal Server Error                   |

## Rate Limiting

API requests are rate limited to ensure fair usage:

* **Standard**: 100 requests per minute
* **Burst**: 10 requests per second

When rate limited, you'll receive a `429` response with retry information.

## SDKs and Libraries

Official SDKs coming soon. In the meantime, use any HTTP client:

<CodeGroup>
  ```javascript Node.js (axios) theme={null}
  const axios = require('axios');

  const client = axios.create({
    baseURL: 'https://api.meetzy.io',
    headers: {
      'Authorization': `Bearer ${process.env.MEETZY_API_TOKEN}`,
      'Content-Type': 'application/json'
    }
  });
  ```

  ```python Python (requests) theme={null}
  import requests

  headers = {
      'Authorization': f'Bearer {API_TOKEN}',
      'Content-Type': 'application/json'
  }

  response = requests.get(
      'https://api.meetzy.io/calls',
      headers=headers
  )
  ```
</CodeGroup>

## Need Help?

<CardGroup>
  <Card title="Create a Call" icon="phone" href="/api-reference/endpoint/create">
    Learn how to initiate calls via API.
  </Card>

  <Card title="Create a Campaign" icon="megaphone" href="/api-reference/endpoint/campaign">
    Launch outbound call campaigns from a contact list.
  </Card>

  <Card title="Get Call Details" icon="info-circle" href="/api-reference/endpoint/get">
    Retrieve detailed call information.
  </Card>

  <Card title="List Calls" icon="list" href="/api-reference/endpoint/list">
    Retrieve your call history.
  </Card>

  <Card title="Support" icon="life-ring" href="mailto:support@meetzy.io">
    Contact our support team.
  </Card>
</CardGroup>


# Advanced Calls View
Source: https://docs.meetzy.es/calls/advanced-view

Explore the new Calls page with a pinnable column manager, predefined and custom saved views, date/agent/company/tag/evaluation/status filters with chips, suspicious-call detection, an OK/KO review verdict filter, summary cards with inbound/outbound transfer time and real AI latency, inline tag editing, per-call audit notes, per-turn latency detail, admin links to Twilio/ElevenLabs, deep links with pre-applied filters, and Excel exports.

The **Advanced Calls View** (`/calls-v2`) is the new version of the Calls page, built for teams that need deeper control over how call data is displayed, filtered, and audited. It introduces a fully configurable column manager (including dynamic input/output columns discovered per agent, right-edge pinning, and a recovered **Evaluations** column), a **Views** menu with both predefined and custom saved views, a two-row filter bar covering date range, agents, company, tags, evaluations, call ID, and status, along with a dedicated suspicious-call filter and an OK/KO verdict filter, a set of summary cards covering volume, latency, review progress, and inbound/outbound transfer time, a manual refresh control, an OK/KO review workflow with per-call audit notes, inline tag editing, per-turn latency detail, admin-only links to Twilio and ElevenLabs, shareable deep links that pre-apply filters via query parameters, full or selective Excel exports, pagination controls at both the top and bottom of the table, and a built-in recording player with a real waveform and previous/next navigation between calls.

<Info>
  The Advanced Calls View is rolling out alongside the classic Calls page. You can switch between the two at any time using the banner at the top of the page.
</Info>

## Switching between views

When you open the Advanced Calls View, you'll see a migration banner at the top of the page confirming you're on the new experience.

* The banner reads **"You're viewing the new Calls page."**, with a **Switch back to the classic view** link right next to it — click it to return to the original Calls page (`/calls-old`) at any time.
* The banner is always visible at the top of the page, so you're never in doubt about which version of the Calls page you're looking at.
* Your filters, column layout, and saved views on the Advanced Calls View are stored independently, so switching back and forth won't affect either page.

<Warning>
  The Advanced Calls View is under active development. If something doesn't behave as expected, use the classic view as a fallback while we continue improving it.
</Warning>

## Configurable and pinnable columns

Unlike the classic Calls page, the Advanced Calls View lets you fully customize which columns appear in the table, in what order, and which ones stay anchored in place. It also includes a dedicated discovery experience for turning any input or output field captured by your agents into a real column.

<Steps>
  <Step title="Open the Columns panel">
    Click the **⚙️ Columns** button in the top-right corner of the page to open the column manager.
  </Step>

  <Step title="Search for a field">
    Use the search box at the top of the panel to look up a specific input or output field by name — for example, typing "lead qualification" filters the discovery panels down to any field whose label or path matches, across every agent, instead of making you scroll through every field manually.
  </Step>

  <Step title="Browse input and output fields by agent">
    Click **Input columns** or **Output columns** to open the corresponding discovery panel — each shows a running count of how many fields were found. Fields are grouped by the **agent** that produced them, with each agent's group collapsed by default; click a group's header to expand or collapse it and reveal the individual fields inside. Opening the input panel automatically closes the output panel (and vice versa), keeping the manager focused on one discovery list at a time.
  </Step>

  <Step title="Add a field as a column">
    Inside an expanded agent group, click **Add** next to any field to turn it into a column immediately. Once added, the button switches to a disabled **Added** state, so you always know at a glance which fields from that agent are already on the table.
  </Step>

  <Step title="Show or hide columns">
    In the **Columns** list below the discovery panels, toggle any column on or off. This includes standard fields (status, duration, agent, company, tags, evaluations) as well as any dynamic input/output columns you've added.
  </Step>

  <Step title="Reorder columns">
    Drag any row in the **Columns** list — using the ⠿ handle — up or down to drop it into the position that makes the most sense for your workflow. The reordering works the same for standard and dynamic columns alike.
  </Step>

  <Step title="Remove a dynamic column">
    Dynamic columns (labeled `input` or `output`) show a small **✕** next to them in the **Columns** list. Click it to remove that column from your layout entirely — this only removes it from view, it doesn't affect the underlying call data, and the field remains available to add again later from the discovery panel.
  </Step>

  <Step title="Pin important columns">
    Click the **pin** icon next to any column to anchor it. Pinned columns are grouped together and stuck to the **right edge** of the table, so they stay visible no matter how many other columns you add or how far you scroll horizontally. **Reviewed** and **Actions** are pinned by default, and you can pin any additional column — such as **Tags** or a dynamic output field you check on every call — the same way. Pinned columns are stacked in order from the right edge inward, and the table automatically measures each one's rendered width so neighboring pinned columns never overlap, even when a pinned column's content varies in length between rows.
  </Step>

  <Step title="Reset to defaults">
    If you want to start over, click **Reset to default** inside the column manager to discard your custom visibility, order, and pinning, restoring the platform's original column layout.
  </Step>
</Steps>

<Info>
  Both discovery panels scroll independently of the field search box and the toggle buttons above them, so you can search or switch between input and output while keeping your place in a long list of agent groups.
</Info>

### Evaluations column

The **Evaluations** column shows which evaluation criteria a call met, mirroring the column that already existed on the classic Calls page.

<Info>
  Evaluations previously existed in the classic view but hadn't been migrated to the Advanced Calls View, so it was missing here and couldn't even be turned on from the column manager. It's now available like any other column and, since it's information many teams audit daily, it's **enabled by default** — including for anyone who already had a saved column layout from before this fix, which is automatically upgraded to include it.
</Info>

### Dynamic columns from inputs and outputs

The Advanced Calls View scans a batch of your loaded calls and automatically discovers structured fields captured on each call's **input parameters** and **agent output**, surfacing them inside the **Input columns** and **Output columns** discovery panels described above, grouped by the agent they came from.

* Every dynamic column is clearly labeled with its **source** — `input` or `output` — so you always know where the value comes from, and which agent it belongs to when relevant.
* Fields are grouped **per agent** in the discovery panel, with each group showing a field count and collapsing by default — handy when several agents each expose dozens of fields and you only need to browse one at a time.
* Use the field search box to jump straight to a known field name without expanding every agent group manually; the search matches against both the field's label and its underlying path.
* Clicking **Add** on a field immediately creates the column and switches the button to **Added**; clicking the **✕** next to that column later removes it from your layout, and it becomes addable again from the discovery panel.
* A handful of internal input fields (like `playbook_id`, `voice_id`, `temperature`, or `record`) are excluded automatically, since they're configuration details rather than meaningful call data.
* Nested objects and arrays are inspected up to a couple of levels deep, so fields buried inside a JSON output (for example inside a `metadata` object) can still surface as their own columns.
* If the same field appears at different paths across calls — for example an agent update that moved a field from the root of the output into a `metadata` object — the column still resolves and displays the value correctly for both older and newer calls, so you don't end up with the value silently missing or duplicated near-identical columns for what's really the same field.

<Info>
  Dynamic fields are re-discovered as new calls load, so new fields introduced by an updated agent version show up as selectable columns automatically, without needing a page refresh or manual configuration.
</Info>

<Warning>
  Resetting columns cannot be undone. If you rely on a custom layout, consider saving it as a view first (see below) before resetting.
</Warning>

## Views

The **👁️ Views** menu in the top-right corner is where you switch between different column-and-filter layouts. It's split into two sections: a set of **predefined views** maintained by Meetzy, and **My views** — the layouts you save yourself. The name of whichever view is currently active is always shown next to the **Views** button, so you know at a glance which layout you're on.

### Predefined views

Predefined views are ready-made layouts covering common review tasks, so you don't have to build a configuration from scratch for everyday use cases:

* **Call audit** — a review-focused layout showing the call date, agent, caller, status, duration, tags, and recording, sorted by most recent first. Ideal for day-to-day quality checks.
* **Latency monitoring** — surfaces the average and maximum AI response latency alongside a latency sparkline, sorted by average latency descending, so the slowest calls float to the top.
* **Compact overview** — a stripped-down layout with just the date, agent, caller, status, and duration, for quickly scanning volume without the extra detail.
* **Transfers & escalations** — highlights transferred and ended calls together with tags and the call summary, useful for reviewing how escalations were handled.

Click any predefined view's name to apply it instantly.

<Info>
  Predefined views can be applied as many times as you like, but they can't be renamed or deleted — they're shared starting points, not personal views. If you want to tweak one, apply it first and then save your adjusted version under a new name in **My views**.
</Info>

### Custom saved views

Once you've configured your columns, pins, and filters the way you like, you can save that configuration as a **named view** and switch back to it whenever you need.

<Steps>
  <Step title="Configure your table">
    Set up your columns, pinning, filters, and sorting exactly how you want them — starting from scratch or from a predefined view.
  </Step>

  <Step title="Open the Views menu">
    Click **👁️ Views** in the top-right corner.
  </Step>

  <Step title="Save the current configuration">
    Type a name for your view in the input field and click **Save current**. Give it something memorable, like "Suspicious calls this week" or "My agent's calls".
  </Step>

  <Step title="Apply a saved view">
    Open the Views menu at any time and click a view's name — predefined or your own — to instantly apply it. The active view is highlighted in green with a checkmark next to it, and its name appears next to the **Views** button. Use the **✕** icon next to one of your own views to delete it.
  </Step>
</Steps>

<Info>
  Saved views are stored per user, so each teammate can build their own set of named views — on top of the same shared predefined views — without affecting anyone else's setup. If you don't have any saved views yet, the **My views** section simply shows a short note letting you know the list is empty.
</Info>

## Filtering calls

The filter bar is organized into two clearly separated rows: a top row holding **Date range**, **Company** (administrator accounts only), **Agents**, **Tags**, **Evaluations**, and **Call ID**, all given equal width so the row stays evenly balanced — and a second, full-width row dedicated entirely to the **Status** chips (plus the **Suspicious calls** and **Review verdict** controls described below). Splitting status onto its own row keeps the chips from wrapping awkwardly next to the other inputs and gives them room to breathe as more statuses (like **Silence**) are added.

### Date range

Choose from common presets (Today, Yesterday, Last 7 days, This month, etc.) or define a fully custom start and end date and time.

<Info>
  Day-based presets here (such as "Today" or "Last 7 days") end at 23:59:59 of the last day, instead of rolling into 00:00 of the following day.
</Info>

### Company

For administrator accounts, a **Company** filter narrows the table down to calls belonging to one or more specific companies — useful when reviewing calls across several accounts at once instead of switching between them one by one.

* Start typing a company name to search, then click it to add it as a chip — you can select several companies at once, the same way the Agents and Tags filters work.
* Remove a chip by clicking the **✕** on it, or clear the whole selection at once using the **✕** at the end of the field.
* Combine it with **Agents**, **Tags**, or **Evaluations** to zero in on, for example, a single agent that belongs to a specific company.
* This filter is only visible to administrator accounts; regular accounts already see calls scoped to their own company and don't need it.

<Info>
  The Company filter reads from the same `playbook_company` data already shown by the optional **Company** column, so the two stay consistent with each other. Selecting a company also narrows the calls available to the **Agents** and **Evaluations** filters accordingly.
</Info>

### Agents

The agent filter uses a chip-based multi-select:

* Start typing an agent's name to search.
* Click an agent from the dropdown to add it as a chip.
* Remove a chip by clicking the **✕** on it, or clear all selected agents at once.
* Leaving the filter empty shows calls from all agents.

### Tags

Tags work the same way as agents — search, select, and manage them as chips. This makes it easy to combine multiple tags to narrow down a specific segment of calls.

* The tag suggestions dropdown is populated from every tag currently used across your calls, so it stays in sync as new tags are created.
* You can combine several tags at once — a call matching **any** of the selected tags is included, making it easy to group related segments (for example, several campaign tags) into a single filtered view.
* Tags added, renamed, or removed through the inline tag editor (see below) are reflected in this filter's suggestions right away.

### Evaluations

The **Evaluations** filter lets you narrow the table down using the same evaluation criteria surfaced in the **Evaluations** column, without having to scan every row manually.

* Filter **by name** to isolate calls that were checked against a specific evaluation criterion — for example, only calls evaluated for "asked for consent" or "confirmed appointment date".
* Filter **by result** to further narrow those calls down to only the ones that **passed**, only the ones that **failed**, or the ones where that evaluation wasn't applicable to the call.
* Combining a criterion name with a result is the fastest way to answer questions like "which calls failed the identity-verification check this week", instead of opening each call individually to check its evaluations.

<Info>
  Like Tags, the list of evaluation names offered in this filter is drawn from the evaluations actually recorded across your calls, so it stays in sync as agents' evaluation criteria evolve.
</Info>

### Call ID

The **Call ID** field lets you jump straight to a single call when you already know its identifier — for example, one referenced in a support ticket or shared by a teammate.

* Paste or type a full or partial call ID to narrow the table down to matching calls.
* This filter combines with the others, so you can also use it to double-check that a specific call falls within a given agent, tag, or date range.
* Clear the field to remove the filter and go back to your broader result set.

### Status

Filter by one or more call statuses, including Open, Closed, No answer, Voicemail, **Silence**, Busy, Failed, Transferred, and Blocked. These chips now sit on their own dedicated row below the other filters, so selecting several statuses at once never crowds the row above.

<Info>
  **Silence** identifies calls where the line connected but no real conversation took place — the call ended without a transcript worth reviewing. It's surfaced as its own selectable status chip, separate from Failed or No answer, so you can isolate these calls specifically instead of digging for them inside a broader status group.
</Info>

### Suspicious calls

The **Suspicious calls** filter is a one-click way to surface calls that are worth a closer look. It's a client-side filter that flags a call as suspicious if **any** of the following are true:

* The AI response latency was unusually high.
* The call ended with a problematic status (`no-answer`, `failed`, or `blocked`).
* The call has no transcript at all.

<Info>
  This is intentionally an **OR** condition rather than an **AND** condition. Combining several individual filters (like status) would only show calls matching *all* conditions, which tends to under-match what auditors actually want to review first. The Suspicious calls filter is designed to catch anything that looks off, from any angle.
</Info>

<Info>
  Because calls with the **Silence** status have no meaningful transcript by definition, they're naturally caught by the "no transcript at all" condition above — so turning on **Suspicious calls** is also a quick way to spot silent calls without setting up a status filter first.
</Info>

The suspicious filter is a single toggle placed alongside the rest of the filter bar — it doesn't require selecting a status or configuring anything else first. Turn it on to instantly narrow the table down to calls worth a closer look, and turn it off to go back to your regular filtered view. It also combines cleanly with the other filters: for example, enabling **Suspicious calls** together with an **Agent** filter shows you only that agent's calls that need attention, without having to build a multi-condition filter manually. A small badge on the toggle reflects how many currently loaded calls match, giving you an immediate sense of scale before you even open the table.

### Review verdict (OK/KO)

A dedicated verdict filter lets you narrow the table down to only the calls your team has already reviewed and marked with an outcome:

* Select **OK** to see everything approved during review.
* Select **KO** to pull up everything flagged as problematic.
* Leave it unset to show calls regardless of verdict, including calls that haven't been reviewed at all.

This is especially handy at the end of a review session — quickly pull up every call marked **KO** to make sure each one has a follow-up note, or check every **OK** call before sharing a "clean" batch with a manager or client. It also combines with the other filters, for example to check the **KO** calls for a single agent, tag, or evaluation result.

## Summary cards

Above the table, a row of summary cards shows aggregate statistics computed **server-side** across **every call matching your current filters and date range** — not just the calls currently loaded on screen. This makes it a reliable, at-a-glance health check even before you scroll through a single row.

The cards include:

* **Volume** — the total number of calls matching your current filters, giving you an immediate sense of scale before diving into the table.
* **Average latency** — the real AI response latency averaged across matching calls, computed from actual turn-by-turn response times rather than an estimate, useful for spotting whether your agents are responding promptly overall.
* **Average duration** — how long matching calls last on average, handy for comparing campaigns or agents against each other.
* **Reviewed** — how many of the matching calls already have a review recorded, so you can track audit progress for a given segment at a glance.
* **Transfer time** — a dedicated card that splits total transfer time into **inbound** and **outbound** totals side by side, so you can see how much time each direction is costing you at a glance, without adding two separate cards to the row.

<Info>
  Because these cards respect your active filters, you can combine them with the **Agents**, **Company**, **Tags**, **Evaluations**, or **Status** filters to check, for example, the average latency for a single agent, or the reviewed count for only the calls tagged as escalations.
</Info>

<Info>
  The transfer time card only accounts for time spent in the transferred portion of a call, and reports inbound and outbound totals separately since the two typically have very different transfer patterns and costs. It's calculated server-side across the full filtered date range, so the numbers stay accurate even when you're only looking at a page of 50 or 100 loaded calls.
</Info>

## Refreshing the table

The table is refreshed **manually**, using a dedicated **↻ Refresh** button near the top of the table.

* Click **Refresh** whenever you want to pull in the newest calls matching your current filters, date range, and sort order.
* Newly arrived calls that weren't in the table before are briefly highlighted, so they're easy to spot right after a refresh.
* Refreshing respects your current filters, sort order, and scroll position — it won't reset or jump you back to the top of the table.

<Info>
  This replaces the previous automatic **Live** polling toggle. Rather than the table quietly refreshing itself in the background at adaptive intervals, you're now fully in control of when the table updates — click **Refresh** whenever you want the latest calls, and the table stays perfectly still in between clicks. This also removes any risk of a background refresh interrupting an in-progress edit, like a review note or tag you're in the middle of typing.
</Info>

<Tabs>
  <Tab title="When to refresh">
    Click **Refresh** right before starting a review session to make sure you're looking at the latest calls, or periodically while monitoring an ongoing campaign to check for new activity — especially combined with the **Suspicious calls** filter to catch anything that needs attention as soon as it shows up.
  </Tab>

  <Tab title="Why manual instead of automatic">
    A manual refresh keeps the table predictable while you work: your scroll position, open editors, and in-progress filters are never disturbed by a refresh you didn't ask for. You decide exactly when new data comes in.
  </Tab>
</Tabs>

## Pagination

To avoid forcing you to scroll all the way down just to move to the next page (and then back up to reach the filters again), pagination controls are shown **twice**: once directly above the table, right next to the refresh control, and again below the table after the last row.

* Both pagination bars stay in sync — changing the page size or moving to another page from either one updates the table and both bars together.
* The top bar is convenient when you've just adjusted a filter and want to immediately move through pages without scrolling.
* The bottom bar is convenient after scanning a full page of results top to bottom, letting you jump to the next page right where you left off.

<Info>
  Both pagination bars always reflect the same total-row count for your current filters, so there's no risk of the top and bottom controls disagreeing on how many pages are available.
</Info>

## Deep links with pre-applied filters

The Advanced Calls View can be opened with filters already applied, using standard URL query parameters. This makes it easy to share a specific, filtered slice of the table with a teammate, or to link out to it from another part of the platform — a dashboard widget, an alert email, or an external report — without asking the recipient to rebuild the filters by hand.

* Supported query params mirror the filter bar itself, including **date range**, **agent**, **company** (for administrators), **tag**, **evaluation** (name and/or result), **status**, **suspicious calls**, and **review verdict**.
* Multiple values for a repeatable filter (such as agents or tags) can be passed together, and are applied as if you had added each one as a chip yourself.
* Opening a deep link populates the filter bar visibly, so you can see exactly which filters were pre-applied and adjust or clear any of them afterward — a deep link is just a starting point, not a locked-down view.
* Deep links combine cleanly with your saved views and column layout: the query-param filters are layered on top of whatever columns and pinning you already have configured, rather than replacing them.

<Info>
  Deep links are especially handy for recurring workflows — for example, bookmarking a link that always opens the table filtered to **Suspicious calls** for a specific agent, or sharing a link to a company's **KO**-verdict calls from the last 7 days directly with that company's account manager.
</Info>

<Warning>
  Query params that reference an agent, company, or tag you don't have access to are simply ignored rather than causing an error, so a mistyped or stale deep link degrades gracefully back to an unfiltered (or partially filtered) view instead of failing to load.
</Warning>

## Review and audit workflow

Each row in the table includes a **Reviewed** control that lets your team track which calls have already been checked, an **OK/KO verdict**, and an optional per-call audit note — all editable directly inline, without opening a separate modal. This column is pinned to the right edge of the table by default, so it stays visible no matter how many columns you've added.

<Steps>
  <Step title="Mark a call as reviewed">
    Click the review indicator on a call's row to toggle its reviewed status inline.
  </Step>

  <Step title="Set an OK/KO verdict">
    Once you've listened to or read through a call, click **OK** or **KO** to record your audit verdict. Clicking the same verdict again clears it. Setting a verdict automatically marks the call as reviewed; clearing the reviewed toggle also clears the verdict.
  </Step>

  <Step title="Add an audit note">
    Click the note icon to open a small inline editor and write context about why the call was flagged, resolved, or worth revisiting — for example, explaining why a call was marked KO. The note is stored per call, so it stays attached to that specific call for anyone reviewing the history later.
  </Step>

  <Step title="Save automatically">
    The reviewed toggle, the verdict, and the note text are all saved automatically a moment after you stop interacting — no need to click a separate save button.
  </Step>
</Steps>

<Info>
  Only one review note editor can be open at a time. Clicking outside the editor, or opening a note on another row, will close the current one.
</Info>

Use the **Review verdict (OK/KO)** filter described above to pull up everything you've already flagged during a session, or to double-check a batch of approvals before sharing them. The **Reviewed** summary card also gives you a quick sense of how much of your current filtered segment still needs attention.

This review workflow is especially useful when combined with the **Suspicious calls** filter: click **Refresh** to catch new calls as they come in, flag the suspicious ones, review them one by one, record an OK/KO verdict with a note explaining the outcome, and mark them as reviewed once resolved. Save that exact combination of filters and columns as a custom view under **My views** — or share it as a deep link — so your team can jump straight back into it later.

<Info>
  This same reviewed toggle, OK/KO verdict, and note editor are also available directly inside the recording player modal, so you can record your audit verdict right after listening to a call without switching back to the table.
</Info>

## Inline tag editing

The **Tags** column lets you manage a call's tags directly from the table, without opening the call's detail view.

<Steps>
  <Step title="Open the tag editor">
    Click into the Tags cell on any row to reveal the inline tag editor.
  </Step>

  <Step title="Add a tag">
    Start typing to search existing tags — an autocomplete dropdown suggests matching tags already used across your calls, so you can stay consistent instead of creating near-duplicate tags. Press **Enter** or select a suggestion to add it.
  </Step>

  <Step title="Rename a tag">
    Renaming isn't a separate action — remove the old tag chip and add the new one in its place. Since tag suggestions are shared across calls, the new name immediately becomes available as a suggestion for every other call going forward.
  </Step>

  <Step title="Remove a tag">
    Click the **✕** on any tag chip to remove it from the call.
  </Step>

  <Step title="Save automatically">
    Tag changes are saved as soon as you add or remove a chip — there's nothing else to confirm.
  </Step>
</Steps>

<Info>
  Tags added or removed inline are immediately reflected in the **Tags** filter's suggestions, so newly created or renamed tags become searchable right away. The same add/remove tag editor is also available inside the recording player modal.
</Info>

## Per-turn latency detail

Beyond the aggregate latency numbers shown in the summary cards, you can drill into a single call to see exactly how its real AI response latency evolved turn by turn.

* Opening a call's latency detail breaks down the response time for each individual agent turn in the conversation, rather than just showing a single average for the whole call.
* This makes it easy to spot whether a high average latency came from one slow turn — for example, a moment where the agent had to call an external tool — or from consistently sluggish responses throughout the call.
* Turn-by-turn latency pairs naturally with the transcript, so you can line up a slow response with exactly what was being discussed at that point in the call.

<Info>
  Use per-turn latency detail together with the **Latency monitoring** predefined view: sort by average latency to find the slowest calls first, then open each one to see precisely which turn is dragging the average up.
</Info>

## Admin links

For users with administrator access, each call includes quick links out to the underlying infrastructure providers, so you don't have to leave the page to cross-reference a call.

* A **Twilio** link opens the corresponding call record directly in the Twilio console, useful for checking carrier-level details like call routing or SIP diagnostics.
* An **ElevenLabs** link opens the corresponding conversation directly in the ElevenLabs dashboard, useful for checking voice generation details or debugging the underlying AI conversation.

<Warning>
  These links, along with the **Company** filter, are only visible to administrator accounts. If you don't see them, your role doesn't currently have access to the underlying provider dashboards or cross-company data.
</Warning>

## Recording player

Clicking a call's recording opens a dedicated player modal built specifically for reviewing audio quickly, with a real waveform rendered from the decoded audio rather than a generic progress bar.

<Tabs>
  <Tab title="Waveform & seeking">
    The waveform is built client-side from the actual audio data using the Web Audio API, so its shape reflects the real volume of the call rather than a placeholder animation. Click anywhere on the waveform to seek directly to that point in the recording.
  </Tab>

  <Tab title="Skip ±10s">
    Use the skip-back and skip-forward controls to jump exactly 10 seconds backward or forward — handy for re-listening to a specific exchange without scrubbing manually.
  </Tab>

  <Tab title="Playback speed">
    Choose a playback speed — 0.5×, 0.75×, 1×, 1.25×, 1.5×, or 2× — to slow down a tricky passage or speed through a long, uneventful call.
  </Tab>
</Tabs>

### Navigating between calls

Instead of closing the modal to move to another row, the player lets you step through calls directly from within it.

<Steps>
  <Step title="Use the arrow buttons">
    Click the **‹** and **›** icons in the modal header to jump to the previous or next call in your current filtered and sorted list.
  </Step>

  <Step title="Or use your keyboard">
    Press the **←** and **→** arrow keys to move between calls without touching the mouse — handy for quickly moving through a batch of flagged calls.
  </Step>

  <Step title="Reach the edges">
    The **‹** and **›** buttons are automatically disabled once you reach the first or last call in the list, so you always know when you've reviewed everything currently loaded.
  </Step>
</Steps>

<Info>
  Moving to the next or previous call keeps the same modal open — the recording, transcript, and review controls all update in place, with a short slide transition indicating which direction you moved.
</Info>

<Info>
  Keyboard shortcuts for navigating between calls are disabled automatically while you're typing in the note or tag input, so arrow keys behave normally while you're writing.
</Info>

<Info>
  The player modal also includes the same **Reviewed**/**OK**/**KO** review controls, per-call audit note, and per-turn latency detail found in the table, plus inline tag management — so you can complete most of a review session, including moving on to the next call, without ever leaving the recording view.
</Info>

## Exporting to Excel

The Advanced Calls View supports two export modes:

<Tabs>
  <Tab title="Export all filtered calls">
    Click **Export to Excel** to export every call currently matching your active filters, respecting your selected columns.
  </Tab>

  <Tab title="Export selected rows only">
    Select one or more rows using the row checkboxes, then click **Export selection (N)**. This exports only the calls you've selected, and always includes the **Reviewed** status, the **OK/KO verdict**, and the **audit note**, regardless of whether those columns are currently visible in the table.
  </Tab>
</Tabs>

<Info>
  Selective export is ideal for sharing a curated set of reviewed or suspicious calls with a manager or client, without exposing your entire filtered dataset. Because the review status, verdict, and audit note are always included in this export, it doubles as a ready-made audit trail you can hand off as-is.
</Info>

<Info>
  The **Export selection (N)** button only appears once you've selected at least one row, with the count updating live as you check or uncheck rows — so you always know exactly how many calls will be included before you export.
</Info>

<Warning>
  Both export modes generate the file client-side based on the calls currently loaded in the table. If you're exporting all filtered calls, make sure you've loaded as many pages as you need before exporting, since the export won't reach beyond what's already been fetched into the table.
</Warning>


# Call Details
Source: https://docs.meetzy.es/calls/call-details

View transcription, recording, input/output data, and more for each call

# Call Details

Every call generates detailed data that helps you understand what happened during the conversation. The Call Details modal provides access to all information captured during the call.

## Opening Call Details

Click on any call row in the Call History table, or click the **View** button to open the details modal.

<Info>
  The recording player described below (waveform, navigation, review, and tags) is a single shared component. It now also powers the recording modal in the **advanced Calls view**, so whichever screen you review calls from, the behavior and shortcuts documented here are identical.
</Info>

## Modal Layout

The Call Details modal is organized in a three-column layout:

| Panel     | Content                 |
| --------- | ----------------------- |
| ⬅️ Left   | Input Parameters (JSON) |
| 📊 Center | Call Information        |
| ➡️ Right  | Output Response (JSON)  |

## Input Parameters

The left panel displays the input JSON that was sent to initiate the call:

```json theme={null}
{
  "phone": "+1234567890",
  "customer_name": "John Smith",
  "appointment_date": "2024-02-15",
  "custom_fields": {
    "order_id": "ORD-12345"
  }
}
```

### Features

* **Copy Button** - Copy the entire JSON to clipboard
* **Syntax Highlighting** - Color-coded JSON for readability
* **Expandable** - Scroll through large input payloads

## Call Information

The center panel shows key metadata about the call:

| Field        | Description                             |
| ------------ | --------------------------------------- |
| Call ID      | Unique identifier (monospace, copyable) |
| Status       | Current status with colored badge       |
| Success      | ✓ Yes or ✗ No indicator                 |
| Playbook     | Agent name that handled the call        |
| Duration     | Call length in human-readable format    |
| From         | Origin phone number                     |
| To           | Destination phone number                |
| Email        | Contact email (if provided)             |
| Created Time | Timestamp when call was initiated       |
| Tags         | Labels applied to this call             |

### Status Badges

| Status      | Color     | Description               |
| ----------- | --------- | ------------------------- |
| completed   | 🟢 Green  | Call finished normally    |
| in-progress | 🔵 Blue   | Call is currently active  |
| failed      | 🔴 Red    | Call encountered an error |
| transferred | 🟣 Purple | Call was transferred      |
| no-answer   | ⚪ Gray    | Recipient didn't answer   |

### Analyze with Copilot

A new **Analyze with Copilot** button is available in the call information section that allows you to dive deeper into the conversation:

<Steps>
  <Step title="Open Copilot Analysis">
    Click the "Analyze with Copilot" button to navigate to the agent editor with this call's context pre-loaded
  </Step>

  <Step title="Automatic Context Loading">
    The system automatically loads the call's transcription and metadata into Copilot for immediate analysis
  </Step>

  <Step title="Ask Questions">
    You can now ask Copilot questions about this specific call, such as:

    * "What were the key topics discussed in this call?"
    * "How could the agent's response be improved?"
    * "What patterns do you see in the customer's behavior?"
  </Step>
</Steps>

<Info>
  The Copilot analysis maintains full context of the call, including transcript, input parameters, and output data for comprehensive insights.
</Info>

## LLM Model Usage

Meetzy runs on ElevenLabs' LLM cascade, which can automatically fall back to a different model mid-call if the primary model is slow or unavailable. The **LLM** block in the Call Information panel gives you visibility into exactly which models handled the conversation.

### Models Used

At the top of the block, you'll see a list of every distinct model that participated in the call, in the order they were first used:

```
gpt-4o-mini → gpt-4o → claude-3-5-sonnet
```

<Info>
  If only one model appears, the cascade never needed to fall back — the primary model handled the entire call. If more than one model appears, ElevenLabs degraded to a secondary model at some point during the conversation.
</Info>

<Warning>
  For calls made before **July 29, 2026**, this list falls back to the single `model_used` field, since per-turn tracking wasn't available yet. In that case, you'll only see one model, even though a cascade may still have been active behind the scenes.
</Warning>

### Per-Turn Breakdown (Cascade Detail)

Below the model list, click **Show Detail** to expand a collapsible, second-by-second breakdown of the cascade. This detail is collapsed by default to keep the modal compact.

Each row in the expanded view represents one conversational turn and shows:

| Column   | Description                                       |
| -------- | ------------------------------------------------- |
| Time     | Elapsed seconds (`t`) since the call started      |
| Model(s) | Which model(s) generated output for that turn     |
| Tokens   | Output tokens produced by each model at that turn |

If a turn used more than one model — for example, the primary model started responding but was interrupted by a fallback — both models are listed together, joined with `+`.

<Steps>
  <Step title="Locate the LLM block">
    Open any completed call's details and find the **LLM** section in the center panel, below the core call metadata.
  </Step>

  <Step title="Review the models used">
    Check the summary list to quickly spot whether the cascade degraded during the call.
  </Step>

  <Step title="Expand the detail">
    Click **Show Detail** to reveal the per-turn table and identify exactly when the switch happened.
  </Step>

  <Step title="Collapse when done">
    Click **Hide Detail** to collapse the breakdown again and keep the modal easy to scan.
  </Step>
</Steps>

<Info>
  Cascade degradation is not necessarily an error — it means ElevenLabs automatically routed the conversation to keep latency low. Use this data to spot patterns (e.g., a model that degrades frequently) and adjust your agent configuration if needed.
</Info>

## Output Response

The right panel displays the output JSON containing all data captured during the call:

```json theme={null}
{
  "transcription": "Agent: Hello, this is Sarah from...",
  "summary": "Customer confirmed appointment for...",
  "sentiment": "positive",
  "output_params": {
    "appointment_confirmed": true,
    "preferred_time": "morning",
    "callback_requested": false
  },
  "evaluations": {
    "successful_greeting": true,
    "collected_requirements": true
  }
}
```

### Output Contents

<AccordionGroup>
  <Accordion title="Transcription">
    Full conversation transcript between the agent and the caller, with speaker labels.
  </Accordion>

  <Accordion title="Summary">
    AI-generated summary of the call's key points and outcomes.
  </Accordion>

  <Accordion title="Output Parameters">
    Structured data extracted from the conversation based on your Output Settings configuration.
  </Accordion>

  <Accordion title="Evaluations">
    Results of automated evaluations defined in your agent configuration.
  </Accordion>

  <Accordion title="Tool Calls">
    If the agent invoked any tools or webhooks, their requests and responses are included.
  </Accordion>
</AccordionGroup>

## Webhook Information

If your agent has a webhook configured, the bottom of the right panel shows:

| Field  | Value                         |
| ------ | ----------------------------- |
| Method | POST, GET, etc.               |
| URL    | Webhook endpoint              |
| Auth   | Authorization header (masked) |

<Info>
  The webhook section only appears if a webhook was configured and triggered for this call.
</Info>

## Recording Playback

If call recording is enabled, clicking the recording icon opens a dedicated **Call Recording** modal with a full-featured player on one side and a rich, per-turn transcript on the other.

<Info>
  This is the same recording modal used across Meetzy — including the **advanced Calls view** — so the waveform player, keyboard navigation, and review/tag editing described below all work identically no matter which list you opened it from.
</Info>

### Navigating Between Calls

The recording modal no longer traps you on a single call. When you open a recording from a filtered list (for example, from Call History), the modal knows about the calls immediately before and after the one you're viewing, and lets you move between them without closing the modal and reopening a different row.

<CardGroup>
  <Card title="Previous / Next buttons" icon="arrow-right-arrow-left">
    Two chevron buttons sit in the modal header, right next to the close button, so they're always visible and never overlap the waveform or transcript below. Click **‹** to load the previous call's recording and transcript, or **›** to load the next one, all without leaving the modal.
  </Card>

  <Card title="Keyboard shortcuts" icon="keyboard">
    Press **←** to go to the previous call and **→** to go to the next call, without touching the mouse. Hovering the header buttons also shows these shortcuts as tooltips ("Previous call (←)" / "Next call (→)").
  </Card>
</CardGroup>

<Info>
  The previous/next buttons are automatically disabled (dimmed, non-clickable) at either end of the list — for example, if you're already viewing the first call, the "previous" button (and the ← shortcut) has no effect. The buttons only render at all when at least one adjacent call is available.
</Info>

<Info>
  Arrow key navigation is automatically disabled while you're typing in the review note or tag input fields inside the modal, so pressing ← or → to move the cursor in a text field never accidentally jumps to another call.
</Info>

Switching calls triggers a short slide transition (the content pane animates left or right depending on the direction you navigated) so it's clear a new recording has loaded, and reloads the waveform, audio, transcript, review status, and tags for the newly selected call.

### Call ID Badge Next to the Transcript

To make it easy to reference a specific call while reviewing recordings — for example, when filing a bug report or pasting the ID into a support ticket — the **Call ID** is now shown as a small badge right above the transcript panel, next to the "Transcript" heading.

<CardGroup>
  <Card title="Always visible" icon="id-badge">
    You no longer need to scroll back up to the Call Information panel or return to the table to find the ID of the call you're currently listening to — it's shown directly next to the transcript.
  </Card>

  <Card title="One-click copy" icon="copy">
    Click the badge to copy the Call ID to your clipboard, exactly like the copy buttons elsewhere in the modal.
  </Card>
</CardGroup>

<Info>
  Because the badge reloads together with the rest of the panel, navigating to the previous/next call with the header buttons or arrow keys automatically updates it to the new call's ID — you'll never accidentally copy the wrong one.
</Info>

### Quick Verdict (Pending / OK / KO)

Reviewing a batch of recordings usually ends with the same decision for each call: was this a good call or not? Instead of closing the modal and marking the call from the table, you can now record that verdict right from the recording player.

The verdict control is **tri-state**, not a simple on/off toggle:

| State   | Meaning                                                                        |
| ------- | ------------------------------------------------------------------------------ |
| Pending | Default state — the call has not been reviewed yet, or the verdict was cleared |
| OK      | The call was handled correctly by the agent                                    |
| KO      | The call needs attention (bad handling, missed data, wrong tone, etc.)         |

<CardGroup>
  <Card title="OK / KO control" icon="thumbs-up">
    A three-way control in the modal, next to the review checkbox, lets you mark the call currently loaded as **OK** or **KO**, with the active option visually highlighted.
  </Card>

  <Card title="Click again to deselect" icon="rotate-left">
    Clicking the currently active option a second time clears it, returning the call to **Pending** — there's no need to hunt for a separate "clear" or "undo" action.
  </Card>
</CardGroup>

<Info>
  The verdict is saved against the call you're viewing and travels with it — navigating to the previous/next call loads that call's own verdict state (Pending, OK, or KO); it never carries over from the call you just left.
</Info>

<Steps>
  <Step title="Open a recording">
    Open any call's recording from Call History or another filtered list. Its verdict starts at whatever state was last saved for it — commonly **Pending** if it hasn't been reviewed yet.
  </Step>

  <Step title="Review the call">
    Listen to the audio and read the transcript to judge whether the agent handled the call well.
  </Step>

  <Step title="Set the verdict">
    Click **OK** or **KO** in the modal to record your judgment for this call. Click the same button again to reset the verdict back to **Pending** if you change your mind.
  </Step>

  <Step title="Move to the next call">
    Use the **›** button or the **→** key to jump straight to the next call in the list and repeat the process — no need to reopen the modal.
  </Step>
</Steps>

<Info>
  Combining the tri-state verdict with previous/next navigation and the arrow-key shortcuts turns batch call review into a fast, keyboard-friendly workflow: listen, judge, move on. The **Pending** state also makes it easy to filter for calls that still need a human decision.
</Info>

<Warning>
  Setting a verdict updates the call's review status immediately. If you need to also leave a written explanation, use the review note field alongside the verdict control — the two are independent and both persist with the call.
</Warning>

### Reviewed Status, Notes, and Tags — Right From the Modal

Full review editing now lives inside the recording modal itself, so you don't need to go back to the table row to finish reviewing a call.

<CardGroup>
  <Card title="Reviewed checkbox" icon="check-double">
    A **Reviewed** checkbox marks the call as looked-at, independent of the OK/KO verdict. Unchecking it also clears the verdict back to Pending, keeping the two fields consistent — the same behavior as the checkbox in the calls table.
  </Card>

  <Card title="Review note" icon="pen">
    A text field lets you type a free-form note explaining your verdict (e.g. "Agent missed the callback time"). It saves automatically a moment after you stop typing, so there's no separate save button to remember.
  </Card>

  <Card title="Tags" icon="tag">
    Add or remove tags for the call directly in the modal footer — press **Enter** after typing a tag to add it, or click the **✕** on an existing tag chip to remove it. Changes apply immediately and are reflected back in the calls table.
  </Card>

  <Card title="Follows the call" icon="arrow-right-arrow-left">
    Just like the verdict, the Reviewed checkbox, note, and tags all reload for whichever call you navigate to with the previous/next controls — nothing is carried over between calls.
  </Card>
</CardGroup>

<Info>
  Because tags and review fields are edited on the same object the table row displays, updates made in the modal show up in the calls list the moment you close it — no manual refresh needed.
</Info>

### Waveform Player

Unlike a plain `<audio>` progress bar, the player renders a **real, interactive waveform** generated directly from the decoded audio — not a placeholder animation. When the recording loads, the audio is decoded via the Web Audio API and sampled into 100 amplitude bars, so the shape you see actually reflects the call's audio (silences, loud segments, etc.).

<CardGroup>
  <Card title="Click-to-seek" icon="hand-pointer">
    Click anywhere on the waveform to jump to that point in the call. The played portion is highlighted in a different color as the call progresses.
  </Card>

  <Card title="+/-10s skip" icon="rotate">
    Use the back/forward buttons to jump exactly 10 seconds, ideal for re-listening to a specific exchange without scrubbing manually.
  </Card>

  <Card title="Playback speed" icon="gauge">
    Choose between 0.5x, 0.75x, 1x, 1.25x, 1.5x, and 2x to review calls faster or slow down a hard-to-understand segment.
  </Card>

  <Card title="Download" icon="arrow-down-to-line">
    Download the original recording file for offline access, sharing, or permanent storage.
  </Card>
</CardGroup>

<Info>
  The waveform is computed once per call and cached for the session — reopening the same recording doesn't require re-decoding the audio.
</Info>

### Enriched Transcript

Next to the player, the transcript is rendered turn-by-turn (not as a single wall of text), with the [Call ID badge](#call-id-badge-next-to-the-transcript) sitting right next to the "Transcript" heading. Each turn shows its elapsed time in the call, and agent turns carry two kinds of expandable detail sourced directly from the underlying ElevenLabs conversation: a **response time breakdown** and, when applicable, **tool call details**.

#### Turn Latency Breakdown

Every agent turn shows a **"Replied in Xs"** summary next to its message. Clicking it expands a per-turn latency breakdown that answers the question auditors actually care about: *the agent took this long to answer — where did that time go?*

The breakdown splits the response time into a stacked bar with the following stages, shown in the order they occur in the pipeline:

| Stage                | Meaning                                                                                           |
| -------------------- | ------------------------------------------------------------------------------------------------- |
| Understanding speech | Speech-to-text (ASR) latency — how long it took to transcribe what the caller said                |
| Model thinking       | Time to first byte from the LLM — how long the language model took to start generating a response |
| Generating voice     | Text-to-speech (TTS) latency — how long it took to synthesize the agent's audio                   |
| Other overhead       | Any small residual not accounted for by the stages above, rather than being silently dropped      |

<Info>
  A separate **"Waiting on purpose"** segment may also appear below the stacked bar. This is not slowness — it's the agent deliberately holding back before speaking (for example, waiting a few seconds before asking "are you still there?" or pausing to let the caller finish a thought). It's called out separately, with an explanatory note, so a deliberate pause isn't mistaken for a performance problem. The headline **"Replied in Xs"** always excludes this deliberate wait time, so it reflects only actual processing latency.
</Info>

Each stage in the breakdown shows its duration and percentage of the total, along with the LLM model that generated the response for that turn. This turns a single opaque latency number into an actionable diagnosis:

* A response dominated by **Understanding speech** may point to background noise or a caller who talks quickly.
* A response dominated by **Model thinking** may point to a slow or overloaded LLM — worth checking against the [LLM Model Usage](#llm-model-usage) section above to see if a cascade fallback happened on that same turn.
* A response dominated by **Generating voice** may point to a longer-than-usual agent reply that takes more time to synthesize.
* A visible **Waiting on purpose** segment simply confirms the agent chose to pause — no action needed.

<Steps>
  <Step title="Scan the turn">
    Each message bubble shows its elapsed time in the call, plus a "Replied in Xs" summary on agent turns.
  </Step>

  <Step title="Expand the latency breakdown">
    Click **Replied in Xs** to reveal the stacked bar showing how the response time splits across understanding speech, model thinking, generating voice, and any deliberate wait.
  </Step>

  <Step title="Check for deliberate waits">
    If a **Waiting on purpose** segment is shown, treat it as intentional agent behavior rather than latency to fix.
  </Step>

  <Step title="Expand tool calls">
    Click a **🔧 tool name** badge to reveal the exact parameters sent to the tool and the result (or error) it returned.
  </Step>
</Steps>

#### Tool Call Details

If the agent invoked a tool or function during a turn, a **🔧 tool name** badge appears next to that turn. Expanding it shows:

* **Parameters** — the exact JSON payload sent to the tool for that call
* **Result** — the value the tool returned, or the error details if the call failed

This lets you verify not just *that* a tool was called, but exactly *what* was sent and *what* came back, without leaving the recording view.

#### Color-Coded Tool Activity

To make tool activity scannable at a glance, the transcript now uses a consistent color code for anything related to tools — both in the badges next to a turn and in the expanded detail cards below it:

| Color    | Meaning                                                                                  |
| -------- | ---------------------------------------------------------------------------------------- |
| 🟠 Amber | A tool call — the request the agent sent to a tool, with the tool name and parameters    |
| 🔵 Blue  | A tool response — the result the tool returned for that call                             |
| 🔴 Red   | A tool error — the call failed, and the card shows the error details instead of a result |

<Info>
  This color coding is applied consistently across the modal: the small **🔧 tool name** badge next to a turn, and the expanded parameters/result card underneath it, both use the same amber/blue/red scheme. A quick glance down the transcript is enough to spot which turns involved tools and whether any of those calls failed.
</Info>

<Info>
  This view mirrors what ElevenLabs' own conversation view shows, since this per-turn detail (latency metrics, model attribution, tool payloads) isn't persisted in Meetzy's database — it's fetched live from ElevenLabs when you open the recording.
</Info>

<Warning>
  If the underlying ElevenLabs conversation is no longer available (for example, an old or purged call), the modal automatically falls back to a **plain transcript** view — the message bubbles are still shown, but without latency breakdowns, model detail, or expandable tool call/result cards. A short notice appears above the transcript in this case. The same amber/blue/red color coding for tool calls, responses, and errors is preserved in this fallback mode wherever that information is still present in the raw transcript, so tool activity remains easy to spot even without the rich per-turn metrics.
</Warning>

<Warning>
  Call recordings are stored for a limited time based on your plan. Download important recordings for permanent storage.
</Warning>

## Enhanced Analytics Workflow

The new Copilot integration streamlines your analysis workflow:

<CardGroup>
  <Card title="Quick Analysis" icon="bolt">
    Jump directly from call details to Copilot analysis without losing context
  </Card>

  <Card title="Contextual Insights" icon="brain">
    Get AI-powered insights based on the specific call's transcription and data
  </Card>

  <Card title="Pattern Recognition" icon="chart-line">
    Identify trends and patterns across similar calls using Copilot's analytical capabilities
  </Card>

  <Card title="Agent Improvement" icon="arrow-up-right">
    Use call-specific analysis to refine agent prompts and behaviors
  </Card>
</CardGroup>

## Copying Data

### Copy Buttons

* **Copy Input** - Copies the entire input JSON
* **Copy Output** - Copies the entire output JSON
* **Copy Call ID** - Copies just the call identifier

### Feedback

When you copy data:

* The button changes to "Copied!" with a checkmark
* Reverts to normal after 2 seconds

## Common Use Cases

<CardGroup>
  <Card title="Debugging" icon="bug">
    Review input params and output to diagnose why a call didn't perform as expected. Use the color-coded tool cards in the transcript to quickly spot which tool call failed.
  </Card>

  <Card title="Quality Assurance" icon="check-double">
    Listen to recordings, check transcriptions, and mark each call Pending/OK/KO — plus add a note and tags — to build a fast, keyboard-driven QA workflow entirely inside the recording modal.
  </Card>

  <Card title="Integration Verification" icon="plug">
    Verify that webhook data was sent correctly and contains expected values.
  </Card>

  <Card title="AI-Powered Analysis" icon="robot">
    Use Copilot to analyze call transcriptions and generate improvement recommendations.
  </Card>
</CardGroup>

## Keyboard Shortcuts

| Shortcut      | Action                                            |
| ------------- | ------------------------------------------------- |
| `Esc`         | Close the modal                                   |
| Click outside | Close the modal                                   |
| `←`           | In the recording modal: jump to the previous call |
| `→`           | In the recording modal: jump to the next call     |

<Info>
  The ← and → shortcuts only apply while the Call Recording modal is open, and are automatically ignored while typing in a text field (such as the review note or tag input).
</Info>

<Info>
  There is no dedicated shortcut for the verdict control, the Reviewed checkbox, or tags — use the on-screen controls in the modal. This keeps review edits a deliberate click/keystroke rather than something that could be triggered accidentally while navigating between calls with the arrow keys. Clicking an already-selected OK or KO button deselects it back to Pending.
</Info>

## Next Steps

<CardGroup>
  <Card title="Call History" icon="list" href="/calls/call-history">
    Return to searching and filtering calls.
  </Card>

  <Card title="Copilot Analysis" icon="brain" href="/agents/copilot">
    Analyze calls with AI-powered insights.
  </Card>

  <Card title="Reporting" icon="chart-line" href="/calls/reporting">
    Generate reports from your call data.
  </Card>
</CardGroup>


# Call History
Source: https://docs.meetzy.es/calls/call-history

Search, filter, and review all your AI agent calls with advanced filtering, auditing, and customization

# Call History

The Call History section provides a comprehensive view of all calls made by your AI agents. You can search, filter, and analyze call data to monitor performance and identify insights with powerful filtering capabilities, an audit/review workflow, and customizable views.

<Info>
  A banner now sits at the very top of this page, letting you switch at any time between the **classic Call History table** and the **new, improved Calls view**. Both views are documented separately — this page covers the improved view; head over to [Call History (Classic)](/calls/call-history-classic) for a full walkthrough of the original table if you're using that view instead.
</Info>

## Accessing Call History

Navigate to **Calls** in the main menu to access your complete call history. This now opens the improved calls view (**Calls v2**) by default.

<Info>
  Calls v2 brings **fully configurable, drag-and-drop columns** (including dynamic input/output fields grouped by agent, with key columns pinned to the right edge), **saved views**, **real AI latency metrics with a per-turn detail chart**, an **expanded filter bar** (date presets, multi-agent selection, tag autocomplete, exact Call ID, latency range, and a one-click "suspicious calls" filter), a built-in **review/audit workflow** with per-call notes, and an **Excel export** that can target either all filtered calls or just your current selection.
</Info>

### Switching Between Views

A notice banner sits at the very top of the page reminding you that you're on the new Calls view.

<Steps>
  <Step title="Notice the Banner">
    At the top of the page, a highlighted banner reads "You're viewing the new Calls page."
  </Step>

  <Step title="Switch Back if Needed">
    Click **Switch back to the classic view** in the banner to be taken to `/calls-old`, the original Call History table, if you run into something that doesn't yet behave as expected in the new view.
  </Step>

  <Step title="Return Anytime">
    Both views read from the same underlying call data, so you can move back and forth between them without losing information — only view-specific preferences like column layout or saved views live separately per view.
  </Step>
</Steps>

<Info>
  This same toggle works in both directions: whether you land here from the main menu or arrive via a direct link, the banner at the top of the page always offers a one-click way back to the classic table, and the classic table offers an equivalent way forward into this improved view. See [Call History (Classic)](/calls/call-history-classic) for the documentation covering that table in detail.
</Info>

<Tip>
  The banner is a safety net during the transition to the new view, not a permanent fixture — treat the classic view as a fallback rather than your primary workflow going forward.
</Tip>

## Interface Overview

The calls interface consists of:

| Section          | Description                                                                                     |
| ---------------- | ----------------------------------------------------------------------------------------------- |
| 🟡 Notice Banner | Reminder that you're on the new view, with a link back to the classic table                     |
| 📊 Header        | Title, description, Views menu, Columns manager, and export actions                             |
| 🔍 Filter Bar    | Advanced filtering controls with visibility management                                          |
| ⚙️ Columns       | Configurable, draggable columns including dynamic agent fields, pinned columns, and saved views |
| 📋 Call Table    | List of all calls with key information, including review status                                 |
| 📄 Details Modal | Detailed view of individual calls, including a full AI latency chart                            |

## Configurable Columns

<Info>
  The improved calls view lets you fully control which columns appear, in what order, and even add columns sourced directly from each agent's input parameters or structured output — no need to open a call individually to see a specific data point.
</Info>

Click the **⚙️ Columns** button above the table to open the column manager panel. It's organized into three areas:

| Section            | Purpose                                                                                                       |
| ------------------ | ------------------------------------------------------------------------------------------------------------- |
| **Input columns**  | Discover fields from each agent's input parameters (`call.data`) and add them as columns, grouped by agent    |
| **Columns**        | Manage visibility and order of all currently active columns — fixed and dynamic — via drag and drop           |
| **Output columns** | Discover fields from each agent's structured output (`call.output`) and add them as columns, grouped by agent |

Saved column layouts live in the separate **👁️ Views** menu next to the Columns button — see [Saved Views](#saved-views) below.

<Steps>
  <Step title="Open the Column Manager">
    Click the **⚙️ Columns** button next to the Export to Excel button to open the column configuration drawer.
  </Step>

  <Step title="Reorder Columns">
    In the **Columns** section, drag any row using the ⠿ handle to reorder it. The table updates its column order live as you drag.
  </Step>

  <Step title="Show or Hide Columns">
    Use the checkbox next to each column name to toggle its visibility. Columns marked as non-toggleable (such as **Actions**) always remain visible.
  </Step>

  <Step title="Add a Dynamic Column from Input or Output">
    Click **Input columns** or **Output columns**, find the field you need under the relevant agent group (grouped and collapsible per agent), and click **Add**. The field appears immediately as a new column, placed just before the pinned columns, and is tagged with an **input**/**output** badge in the Columns list.
  </Step>

  <Step title="Remove a Dynamic Column">
    Dynamic columns display a small ✕ button next to their badge in the Columns list. Click ✕ to remove them from your view entirely.
  </Step>

  <Step title="Reset to Defaults">
    Click **Reset to default columns** at the bottom of the panel to discard all customizations and restore the original column set.
  </Step>
</Steps>

<Tip>
  Your column configuration — order, visibility, and any dynamic columns you've added — is saved automatically per user and restored the next time you open the view. There's no manual save step.
</Tip>

<Warning>
  Dynamic columns are tied to the agent and field path they were created from. If an agent stops sending a particular input or output field, that column will simply show empty values for calls where the data is missing, rather than being removed automatically.
</Warning>

### Pinned Columns

<Info>
  A small number of columns are **pinned to the right edge** of the table so they stay visible regardless of how many columns you add or how far you scroll horizontally.
</Info>

By default, **Reviewed** and **Actions** are pinned. This keeps your audit status and row-level actions within reach even when you've added several dynamic input/output columns that push the table wide.

<Tip>
  Pinned columns are visually grouped together at the right edge — if you rely heavily on the review workflow, keep the **Reviewed** column pinned rather than hiding it, so you always know which calls in view still need attention.
</Tip>

### Saved Views

<Info>
  Beyond your single auto-saved layout, the **👁️ Views** menu (next to the Columns button in the header) lets you keep multiple **named views** — for example, one focused on latency troubleshooting and another focused on audit review — and switch between them instantly.
</Info>

<Steps>
  <Step title="Configure a Layout">
    Show, hide, reorder, and add dynamic columns until the table matches the view you want to save.
  </Step>

  <Step title="Open the Views Menu">
    Click **👁️ Views** in the header to open the saved views panel.
  </Step>

  <Step title="Name and Save">
    Type a name into the **New view name…** field and click **Save current**. The current column configuration is stored as a new saved view.
  </Step>

  <Step title="Apply a Saved View">
    Click on any saved view's name in the list to instantly replace your current column configuration with that view's layout.
  </Step>

  <Step title="Delete a Saved View">
    Click the ✕ next to a saved view to remove it. This only deletes the saved view itself — it doesn't change your currently active columns.
  </Step>
</Steps>

<Tip>
  Saved views only capture the **column configuration** (which columns, their order, and visibility) — they don't store filters, sorting, or page size. Combine a saved view with a filter preset for a fully repeatable workflow.
</Tip>

### Legacy Column Configuration Migration

<Info>
  If you previously customized columns in the classic Call History table, that layout isn't lost — the improved view automatically imports it the first time you open it.
</Info>

The first time a user opens the improved calls view without any existing saved preferences, the system performs a **one-time, automatic migration**:

| Step | What happens                                                                                                                          |
| ---- | ------------------------------------------------------------------------------------------------------------------------------------- |
| 1    | The improved view checks your browser's local storage for a legacy column configuration saved from the classic Call History table     |
| 2    | Fixed columns that exist in both the old and new default sets carry over their visibility setting                                     |
| 3    | Any dynamic input or output columns you had added in the classic table are reconstructed as-is, keeping their label, path, and source |
| 4    | Your previous sort column, sort direction, and page size are applied if present                                                       |
| 5    | The migrated configuration is saved to your account immediately, so it's available even if you switch devices or browsers afterward   |

<Warning>
  This migration is read-only and runs only once. It reads from the browser's local storage, so it only applies on the same browser/device where you originally customized the classic table. If you've already opened the improved view before, or already have saved preferences for it, the migration is skipped and your existing improved-view configuration is used instead.
</Warning>

### Default Column Set

Out of the box, the improved view includes:

| Column              | Description                                                      |
| ------------------- | ---------------------------------------------------------------- |
| Date                | When the call occurred                                           |
| Agent               | Name of the agent that handled the call                          |
| Agent ID            | The agent's unique identifier, shown as a compact copyable badge |
| Call ID             | The call's unique identifier, shown as a compact copyable badge  |
| From                | Caller phone number                                              |
| Status              | Call status badge                                                |
| Goal                | Success indicator based on evaluations                           |
| Duration            | Call length                                                      |
| Avg AI latency      | Average AI response latency for the call                         |
| Latency (turns)     | Compact sparkline of latency across the call's turns             |
| Tags                | Dynamic tags associated with the call                            |
| Recording           | Access to the call recording player                              |
| Reviewed *(pinned)* | Audit/review status for the call, with an inline note editor     |
| Actions *(pinned)*  | Row-level quick actions                                          |

Additional columns such as **To**, **Max latency**, **Company**, **Email**, **Summary**, **Type**, **Hung up by**, and **Agent duration** are available but hidden by default — enable them from the column manager as needed.

<Tip>
  The **Agent ID** and **Call ID** columns render as compact badges showing only the last 10 characters with a one-click copy button — handy for grabbing an exact ID to paste into the Call ID filter without cluttering the row with a full UUID.
</Tip>

## Review & Audit Workflow

<Info>
  Every call can now be marked as reviewed and annotated with a free-text note directly from the table — no need to leave Call History to track your QA process.
</Info>

The **Reviewed** column (pinned to the right by default) is the entry point for the audit workflow:

| Element             | Description                                                                                                        |
| ------------------- | ------------------------------------------------------------------------------------------------------------------ |
| **Reviewed toggle** | Marks the call as reviewed or not yet reviewed, visually distinguished in the column                               |
| **Note popover**    | An inline editor attached to the Reviewed column where you can write a free-text audit note for that specific call |
| **Auto-save**       | Notes are saved automatically shortly after you stop typing, without a separate save button                        |

<Steps>
  <Step title="Open the Review Popover">
    Click the **Reviewed** cell for any call to open its inline note editor.
  </Step>

  <Step title="Mark as Reviewed">
    Toggle the call's reviewed status to indicate it has been audited.
  </Step>

  <Step title="Write an Audit Note">
    Type your observations, findings, or QA comments directly into the note field. Your input is debounced and saved automatically a moment after you stop typing.
  </Step>

  <Step title="Close the Popover">
    Click anywhere outside the popover to close it. Only one review popover can be open at a time, so opening a new one automatically closes any other that was active.
  </Step>

  <Step title="Revisit Later">
    Reopen the Reviewed cell for any call at any time to read or edit the existing note — your review history for that call is preserved.
  </Step>
</Steps>

<Tip>
  Combine the Reviewed column with a saved filter preset (for example "Failed + Last 7 Days") and a saved column view that keeps **Reviewed** and **Tags** visible, to build a repeatable QA sweep: filter down to the calls you need to audit, review each one, leave a note, and move to the next.
</Tip>

<Warning>
  Review status and notes are stored per call and are visible to anyone with access to Call History — treat audit notes as shared team documentation rather than private scratch notes.
</Warning>

## AI Latency Metrics

<Info>
  Every call now tracks how quickly your AI agent responded throughout the conversation, giving you a direct signal of conversational responsiveness independent of call outcome or duration.
</Info>

The improved view surfaces AI latency at three levels of detail:

| Metric                 | Column          | Description                                                                                          |
| ---------------------- | --------------- | ---------------------------------------------------------------------------------------------------- |
| **Average latency**    | Avg AI latency  | The mean AI response latency (in milliseconds) across all turns in the call                          |
| **Maximum latency**    | Max latency     | The single highest response latency recorded during the call — useful for spotting worst-case delays |
| **Per-turn sparkline** | Latency (turns) | A compact inline chart showing how latency evolved turn by turn, without leaving the table           |

<Steps>
  <Step title="Scan Averages at a Glance">
    Sort the table by **Avg AI latency** to quickly identify calls or agents with consistently slow responses.
  </Step>

  <Step title="Spot Outliers with Max Latency">
    Enable the **Max latency** column (hidden by default) to catch calls with an isolated but severe delay that the average might hide.
  </Step>

  <Step title="Read the Sparkline">
    The **Latency (turns)** column renders a lightweight line for each call, plotting latency per turn. A rising line toward the end of a call can indicate the agent slowing down as the conversation got more complex.
  </Step>

  <Step title="Open the Full Latency Detail Modal">
    Click into a call's details to open the dedicated AI latency modal: a full, interactive chart plotting latency per turn, with exact millisecond values on hover and the timestamp of each turn along the X axis.
  </Step>
</Steps>

### Latency Detail Modal

<Info>
  The latency detail modal turns the compact table sparkline into a full-size, turn-by-turn chart so you can pinpoint exactly where a call slowed down.
</Info>

| Feature                      | Description                                                                                                                      |
| ---------------------------- | -------------------------------------------------------------------------------------------------------------------------------- |
| **Per-turn chart**           | Every AI turn in the call is plotted as a point, connected in chronological order                                                |
| **Exact values on hover**    | Hovering a point shows the precise latency in milliseconds for that specific turn                                                |
| **Timestamp axis**           | The X axis reflects when each turn occurred during the call, making it easy to correlate spikes with a stage of the conversation |
| **Visual outlier detection** | Sudden jumps in the line stand out immediately compared to scanning raw numbers in a table                                       |

<Tip>
  Combine latency metrics with the LLM and LLM detail columns to check whether latency spikes correlate with a model cascade fallback — a secondary model taking over mid-call can sometimes explain a sudden increase in response time. Open the latency detail modal and the LLM detail breakdown side by side for the same call to confirm.
</Tip>

<Warning>
  Latency data depends on the underlying voice provider reporting response timing for each turn. Calls made before latency tracking was enabled, or through integrations that don't report this data, will show empty values for these columns and an empty chart in the detail modal.
</Warning>

## Recording Player with Waveform

<Info>
  The recording player has been upgraded from a simple audio control to a full waveform-based player, built from the actual decoded audio of the call — not a generic placeholder animation. It also includes a rich, per-turn conversational view when detailed provider data is available.
</Info>

Click the **Recording** column on any call to open the player modal, which includes:

| Feature                    | Description                                                                                                                          |
| -------------------------- | ------------------------------------------------------------------------------------------------------------------------------------ |
| **Real waveform**          | Rendered directly from the decoded audio, showing actual amplitude across the call                                                   |
| **Click-to-seek**          | Click anywhere on the waveform to jump to that point in the recording                                                                |
| **Playback progress**      | The waveform visually fills in as the recording plays, distinguishing played from unplayed audio                                     |
| **Skip controls**          | Jump backward or forward 10 seconds with dedicated buttons                                                                           |
| **Playback speed**         | Choose from 0.5x, 0.75x, 1x, 1.25x, 1.5x, or 2x speed                                                                                |
| **Rich synced transcript** | Per-turn transcript with latency badges (ASR, TTS, LLM), the LLM model used, and expandable tool call/result details, when available |
| **Fallback transcript**    | Falls back to a plain text transcript automatically if detailed per-turn provider data isn't available for that call                 |
| **Download**               | Download the recording file directly from the modal                                                                                  |

<Steps>
  <Step title="Open the Player">
    Click the Recording column for any call with an available recording.
  </Step>

  <Step title="Review the Waveform">
    Wait for the audio to load and decode — the waveform renders automatically once ready, reflecting the real amplitude of the call.
  </Step>

  <Step title="Seek and Adjust Speed">
    Click on any part of the waveform to jump there instantly, or use the speed buttons to review a section more quickly or more carefully.
  </Step>

  <Step title="Cross-Reference with the Transcript">
    Use the transcript panel next to the player to follow along with what's being said as you scrub through the recording. Expand latency or tool badges on any turn for a deeper technical breakdown of that moment in the call.
  </Step>
</Steps>

<Tip>
  Because the waveform is built from decoded audio rather than a static image, longer recordings may take a moment to render on first open. Once decoded, seeking and speed changes are instant.
</Tip>

## Excel Export

<Info>
  Exporting to Excel respects exactly what you see: only the columns currently visible in your table — in the order you've arranged them — are included in the exported file, including any dynamic input/output columns you've added and the current review status/notes.
</Info>

Two export options are available side by side in the header:

| Button                   | Scope                                                                                |
| ------------------------ | ------------------------------------------------------------------------------------ |
| **Export to Excel**      | Exports every call matching your current filters, across all pages                   |
| **Export selection (N)** | Appears only once you've checked one or more rows; exports only those selected calls |

<Steps>
  <Step title="Configure Your View">
    Use the column manager to show, hide, and reorder columns until the table reflects exactly the data you want to export. Apply a saved view first if you have one ready for this purpose.
  </Step>

  <Step title="Apply Your Filters">
    Narrow down the calls you want using the filter bar, since the full export follows the currently filtered result set.
  </Step>

  <Step title="Choose Full or Selective Export">
    Click **Export to Excel** to export every filtered call, or select specific rows with the checkboxes and click **Export selection (N)** to export only those calls — useful when you've hand-picked a set of calls during a review session.
  </Step>

  <Step title="Watch the Export State">
    The button shows an "Exporting…" state while the file is being generated.
  </Step>

  <Step title="Open the File">
    The downloaded spreadsheet mirrors your visible columns and their order, including any dynamic agent input/output fields you've added and the Reviewed status/notes, making it easy to share exactly the view you were working with.
  </Step>
</Steps>

<Tip>
  "Export selection" is a natural companion to the review workflow: mark and annotate the calls you've audited, select them, and export just that batch as a record of your QA pass — without pulling in every other call matching your filters.
</Tip>

<Warning>
  If you hide a column after exporting, previously downloaded files aren't affected retroactively — each export is a snapshot of your column configuration, filters, and (for selective exports) selection at that moment.
</Warning>

## Advanced Filter System

### Filter Visibility Configuration

<Info>
  Customize which filters are visible in your interface to streamline your workflow and focus on the most relevant filtering options.
</Info>

A dedicated **⚙️ settings button** is located at the end of the Filter Bar, next to the filter controls. It opens a configuration panel that lets you show or hide individual filters without affecting your applied filter values.

<Steps>
  <Step title="Open Filter Settings">
    Click the ⚙️ settings icon at the right end of the Filter Bar to open the filter visibility panel.
  </Step>

  <Step title="Toggle Individual Filters">
    Use the checkboxes to show or hide each of the following filters independently:

    * **Goal** — filter calls by the agent's configured goal or objective
    * **Tags** — filter by dynamic or custom tags
    * **Evaluations** — filter by evaluation completion or outcome
    * **Call ID** — the direct database lookup search field, matched on the exact ID

    Core filters such as Date Range, Agents, Status, Direction, Success, Latency range, and Suspicious calls are always visible and cannot be hidden, since they are considered essential for everyday use.
  </Step>

  <Step title="Apply and Close">
    Close the panel by clicking outside of it or pressing the settings icon again. Hidden filters are immediately removed from the Filter Bar, freeing up space for the ones you actually use.
  </Step>

  <Step title="Automatic Persistence">
    Your filter visibility preferences are saved automatically per user and restored the next time you open Call History — no need to reconfigure them on every visit.
  </Step>
</Steps>

<Info>
  By default, the **Tags** filter is visible in the Filter Bar out of the box, so you don't need to open the settings panel to start filtering calls by tags. Goal, Evaluations, and Call ID remain part of the same visibility configuration and follow their own default state, which you can adjust at any time.
</Info>

<Tip>
  If a filter is hidden but has an active value from a previous session, it will remain applied to your results even though it's not visible. Re-enable the filter from the settings panel to review or clear it. Look for the 👁️ **Hidden Filter** indicator near the settings icon — it appears whenever a hidden filter still holds an active value, so you always know your results are being narrowed by something you can't currently see.
</Tip>

### Date Range Filter

<Info>
  The date range filter now offers a much wider set of quick presets, from short rolling windows measured in minutes to full calendar months, alongside a fully custom range with time precision.
</Info>

Click the date range field to open the preset picker:

| Preset          | Range covered                                       |
| --------------- | --------------------------------------------------- |
| Last 15 minutes | Rolling 15-minute window ending now                 |
| Last 30 minutes | Rolling 30-minute window ending now                 |
| Last 1 hour     | Rolling 1-hour window ending now                    |
| Last 2 hours    | Rolling 2-hour window ending now                    |
| Last 4 hours    | Rolling 4-hour window ending now                    |
| Last 8 hours    | Rolling 8-hour window ending now                    |
| Today           | Start of today through end of today                 |
| Yesterday       | Start of yesterday through end of yesterday         |
| Last 7 days     | 7 days ago through end of today                     |
| Last 30 days    | 30 days ago through end of today                    |
| This month      | Start of the current calendar month through its end |
| Last month      | The entire previous calendar month                  |
| Custom range    | Manually chosen start and end date/time             |

<Steps>
  <Step title="Open the Picker">
    Click the date range field in the Filter Bar to reveal the list of presets.
  </Step>

  <Step title="Pick a Preset">
    Click any preset to apply it instantly — the field updates to reflect your selection and the table refreshes.
  </Step>

  <Step title="Or Set a Custom Range">
    Scroll to the **Custom range** section at the bottom of the picker and set an exact **From** and **To** date and time using the datetime inputs.
  </Step>

  <Step title="Confirm">
    Click **Done** to close the picker once your custom range is set.
  </Step>
</Steps>

<Info>
  Use the datetime picker to set precise time ranges for your searches. The filter automatically adjusts to your local timezone.
</Info>

<Tip>
  The short rolling-window presets (15 minutes up to 8 hours) are especially useful right after a batch of outbound calls finishes, letting you review the very latest results without having to construct a custom range.
</Tip>

### Multi-Select Agent Filter with Searchable Combobox

<Info>
  The Agent filter is a typeahead combobox that supports selecting **multiple agents at once**, so it stays fast and usable even when your account has hundreds of agents and you need to compare a handful of them together.
</Info>

<Steps>
  <Step title="Start Typing">
    Click into the Agent field and type any part of the agent's name. As soon as the field is focused, a dropdown of matching agents appears.
  </Step>

  <Step title="Filter as You Type">
    The list narrows in real time to agents whose name contains what you've typed, capped to the closest matches so the dropdown stays responsive.
  </Step>

  <Step title="Select One or More Agents">
    Click an agent from the list to add it as a chip inside the field. Keep typing and selecting to add additional agents — each selected agent appears as its own removable chip.
  </Step>

  <Step title="Remove a Single Agent">
    Click the ✕ on any chip to remove just that agent from the selection, keeping the rest of your filter intact.
  </Step>

  <Step title="Clear All Agents">
    Click the ✕ at the end of the field to remove every selected agent at once and return to viewing calls from all agents.
  </Step>
</Steps>

<Tip>
  When filtering by a single agent, the **"View Evaluation Trends"** button appears next to the agent chip for quick access to performance analytics.
</Tip>

### Status Filter

Filter by call status with visual indicators:

| Status         | Description                | Indicator    |
| -------------- | -------------------------- | ------------ |
| 🟢 Completed   | Call finished successfully | Green badge  |
| 🟡 In Progress | Call currently active      | Yellow badge |
| 🔴 Failed      | Call encountered an error  | Red badge    |
| ⚪ No Answer    | Recipient didn't answer    | White badge  |
| 🔵 Transferred | Call was transferred       | Blue badge   |

### Direction Filter

* **Inbound** - Calls received by your agents with 📞 icon
* **Outbound** - Calls initiated by your agents with 📲 icon
* **All** - Both directions

### Success Filter

Filter by call outcome with enhanced evaluation criteria:

* **Successful** - Calls marked as successful based on your evaluations
* **Unsuccessful** - Calls that didn't meet success criteria
* **All** - All calls regardless of outcome

### Goal Filter

<Info>
  The Goal filter lets you narrow down calls by the objective configured on the agent, making it easier to review calls tied to a specific business intent (e.g., booking, qualification, support resolution).
</Info>

This filter can be shown or hidden from the filter visibility panel described above. It is hidden by default; enable it from the ⚙️ settings panel if you need to filter calls by their configured goal.

Like every optional filter, Goal is temporarily disabled the moment you start a Call ID search — see below for why.

### Latency Range Filter

<Info>
  Filter calls by their average AI response latency to isolate calls that were noticeably fast or slow, without needing to open the latency detail modal one call at a time.
</Info>

| Field           | Description                                                                          |
| --------------- | ------------------------------------------------------------------------------------ |
| **Min latency** | Only show calls whose average AI latency is at or above this value (in milliseconds) |
| **Max latency** | Only show calls whose average AI latency is at or below this value (in milliseconds) |

<Steps>
  <Step title="Set a Minimum">
    Enter a value in **Min latency** to exclude calls that responded faster than that threshold.
  </Step>

  <Step title="Set a Maximum">
    Enter a value in **Max latency** to exclude calls that responded slower than that threshold.
  </Step>

  <Step title="Combine Both">
    Use both fields together to isolate a specific latency band, for example calls between 1500 ms and 3000 ms, to investigate a particular tier of responsiveness.
  </Step>

  <Step title="Clear the Range">
    Empty either field to remove that bound and widen the results again.
  </Step>
</Steps>

<Tip>
  Pair the latency range filter with the **Avg AI latency** and **Latency (turns)** columns so you can immediately see, row by row, why each call matched your threshold.
</Tip>

### Suspicious Calls Filter

<Info>
  The "Suspicious calls" filter is a one-click way to surface calls that are more likely to need a closer look, without manually combining several filters yourself.
</Info>

Enabling this filter highlights calls that match **any** of the following conditions:

| Condition            | Why it's flagged                                                              |
| -------------------- | ----------------------------------------------------------------------------- |
| High average latency | The call's average AI response latency exceeds a built-in threshold (3000 ms) |
| Problematic status   | The call ended as **no-answer**, **failed**, or **blocked**                   |
| Goal not met         | The call's evaluations indicate its configured goal wasn't achieved           |
| Missing transcript   | No transcript was captured for the call at all                                |

<Steps>
  <Step title="Enable the Filter">
    Toggle **Suspicious calls** on in the Filter Bar.
  </Step>

  <Step title="Review the Narrowed Results">
    The table updates to show only calls matching at least one of the suspicious conditions above.
  </Step>

  <Step title="Investigate Each Call">
    Open each flagged call's details, recording, or latency chart to determine whether it needs a review note, a tag, or escalation.
  </Step>

  <Step title="Mark as Reviewed">
    Use the **Reviewed** column to mark calls you've checked and leave an audit note, so the next pass through your suspicious calls doesn't re-cover the same ground.
  </Step>
</Steps>

<Info>
  Unlike Min/Max latency or Status, which apply a strict AND to your other filters, "Suspicious calls" applies an **OR** across its own internal conditions — a call needs to match only one of them to show up, since each condition on its own is a reasonable independent signal of a problematic call.
</Info>

<Tip>
  Combine "Suspicious calls" with a date range like **Today** or **Last 24 hours** to build a quick daily triage habit: check what's suspicious since yesterday, review it, and clear the queue.
</Tip>

### Exact Call ID Search

<Warning>
  When using Call ID search, all other filters are temporarily disabled to ensure precise lookups.
</Warning>

The Call ID filter performs a **direct, exact-match** database lookup and has **exclusive priority over every other filter** in the Filter Bar. As soon as you start typing a value into the Call ID field, all remaining filters (Date Range, Agents, Status, Direction, Success, Goal, Tags, Evaluations, Latency range, and Suspicious calls) become visually disabled and are excluded from the query — the interface greys them out and blocks interaction with them to make it clear that only the Call ID lookup is active.

<Info>
  Think of it as a small measurement problem: the moment you observe the system through a Call ID, every other possible filter state "collapses" — much like a quantum measurement collapsing a wave function into a single outcome. Before you type an ID, the Filter Bar exists in a superposition of Date Range, Agents, Status, Direction, Success, Goal, Tags, Evaluations, Latency range, and Suspicious calls, all potentially active at once. The instant a Call ID value is present, that superposition resolves into a single, deterministic state: exactly one call, found by its exact ID, with no other filter contributing to the result.
</Info>

Enhanced Call ID functionality with database search:

<Steps>
  <Step title="Enter Call ID">
    Type the specific call ID in the dedicated search field. You can also copy an ID directly from the **Call ID** column's compact badge (or the row's Copy ID quick action) and paste it here. As soon as a value is present, the rest of the filters switch to a disabled visual state.
  </Step>

  <Step title="Database Query">
    The system performs a direct database lookup for the exact call ID, ignoring any values currently set in other filters. Partial or fuzzy matches are not returned — the ID must match exactly.
  </Step>

  <Step title="Instant Results">
    The matching call is displayed immediately, bypassing all other filters.
  </Step>

  <Step title="Clear Search">
    Remove the call ID to restore normal filtering functionality. All previously configured filters are automatically re-enabled and reapplied — the "superposition" is restored exactly as you left it.
  </Step>
</Steps>

<Tip>
  Like the Goal and Evaluations filters, the Call ID field is hidden from the Filter Bar by default and can be enabled from the filter visibility settings (⚙️) whenever your workflow requires manual, exact ID lookups.
</Tip>

<Warning>
  This priority behavior is exclusive and non-negotiable: while a Call ID value is present, there is no way to combine it with Date Range, Agents, Status, Direction, Success, Goal, Tags, Evaluations, Latency range, or Suspicious calls. If you need a combined search, clear the Call ID field first and use the other filters instead.
</Warning>

### Dynamic Tags Filter with Autocomplete

<Info>
  Filter calls by dynamic tags that are automatically generated based on call content, outcomes, and agent behaviors, using a chip-based multi-select field with live autocomplete.
</Info>

<Steps>
  <Step title="Start Typing a Tag">
    Click into the Tags field and start typing. A dropdown appears listing existing tags that match what you've typed.
  </Step>

  <Step title="Select a Suggested Tag">
    Click a suggestion to add it as a chip in the field. You can add multiple tags this way to filter for calls matching any of them.
  </Step>

  <Step title="Remove a Tag">
    Click the ✕ on any tag chip to remove just that tag, or clear the field entirely to remove all tag filters at once.
  </Step>
</Steps>

The tags filter provides:

* **Dynamic Tag Generation** - Tags are automatically created based on call analysis
* **Autocomplete Suggestions** - Existing tags are suggested as you type, so you don't need to remember exact spelling
* **Multi-Select Capability** - Select multiple tags to find calls matching any or all criteria
* **Real-time Updates** - Tag list updates as new calls are processed
* **Custom Tag Support** - Manual tags can be added to calls for additional categorization

<Info>
  The Tags filter is **visible by default** in the Filter Bar, so it's ready to use as soon as you open Call History. You can still hide it from the ⚙️ filter settings panel if you prefer a more minimal filter set. Like all other filters, it is temporarily disabled while a Call ID search is active.
</Info>

### Evaluations Filter

<Info>
  Filter calls based on evaluation status, letting you quickly locate calls that are pending review, partially evaluated, or fully completed.
</Info>

Available options include:

* **Completed** - Calls with all evaluations processed
* **Pending** - Calls waiting for evaluation
* **Partial** - Calls with some but not all evaluations completed

This filter is hidden by default in a minimal setup and can be enabled from the filter visibility panel (⚙️) whenever you need deeper evaluation-based analysis. It is also disabled while a Call ID search is active, following the same priority behavior as the rest of the filters.

### Filter Presets

<Tip>
  Save your most-used filter combinations as presets for quick access to common search patterns.
</Tip>

Create and manage filter presets:

<Steps>
  <Step title="Configure Filters">
    Set up your desired combination of filters (date range, agents, status, latency range, suspicious calls, etc.)
  </Step>

  <Step title="Save as Preset">
    Click "Save Preset" and give it a descriptive name
  </Step>

  <Step title="Quick Access">
    Access saved presets from the preset dropdown for instant filter application
  </Step>

  <Step title="Manage Presets">
    Edit or delete presets as your filtering needs evolve
  </Step>
</Steps>

<Tip>
  Filter presets and column saved views are independent: a filter preset remembers *which calls* you're looking at, while a saved view remembers *which columns* you see. Pair a specific preset with a specific saved view for a fully repeatable workflow — for example, "Suspicious + Last 7 Days" alongside a "Review & audit" column layout that keeps Reviewed pinned and visible.
</Tip>

## Enhanced Visual Indicators

### Filter Status Indicators

The interface now includes improved visual feedback:

| Indicator         | Meaning                                                                                                            |
| ----------------- | ------------------------------------------------------------------------------------------------------------------ |
| 🟦 Blue Badge     | Active filter with values selected                                                                                 |
| 🔢 Count Badge    | Number of items selected in multi-select filters (agents, tags)                                                    |
| 🚫 Disabled State | Filter temporarily disabled (e.g., during Call ID search)                                                          |
| ⚙️ Settings Icon  | Opens the filter visibility configuration panel (controls Goal, Tags, Evaluations, and Call ID)                    |
| 👁️ Hidden Filter | Filter is configured as not visible in the current view, but may still hold an active value affecting your results |
| ✨ Preset Active   | A saved preset is currently applied                                                                                |

<Tip>
  The 👁️ Hidden Filter indicator is especially useful right after enabling the Call ID search: once you clear it, glance at the ⚙️ panel to confirm which of your previously hidden filters silently came back into play.
</Tip>

### Selection Status

When calls are selected:

* **Selection Counter** - Shows "Export selection (N)" in the header once at least one call is checked
* **Bulk Action Bar** - Appears with available actions for selected calls
* **Clear Selection** - One-click option to deselect all calls

## Call Table

The call table displays key information for each call with enhanced visual design:

| Column                                  | Description                                                                  |
| --------------------------------------- | ---------------------------------------------------------------------------- |
| Selection                               | Checkbox to select calls                                                     |
| Status                                  | Call status badge with color coding                                          |
| Success                                 | ✓ or ✗ indicator with evaluation context                                     |
| Agent                                   | Name of the agent that handled the call                                      |
| Agent ID *(optional)*                   | The agent's ID, shown as a compact copyable badge                            |
| Call ID *(optional)*                    | The call's ID, shown as a compact copyable badge                             |
| Direction                               | 📞 Inbound or 📲 Outbound with clear icons                                   |
| From                                    | Caller phone number with formatting                                          |
| To                                      | Recipient phone number with formatting                                       |
| Duration                                | Call length in minutes:seconds                                               |
| Date                                    | When the call occurred with relative time                                    |
| Tags                                    | Dynamic tags associated with the call                                        |
| LLM *(optional)*                        | Language model(s) that handled the call                                      |
| LLM detail *(optional)*                 | Turn-by-turn breakdown of model usage                                        |
| Avg AI latency *(optional)*             | Mean AI response latency for the call                                        |
| Max latency *(optional)*                | Highest AI response latency recorded in the call                             |
| Latency (turns) *(optional)*            | Sparkline of latency per turn                                                |
| Reviewed *(pinned)*                     | Audit/review status with an inline note editor                               |
| *(any input/output field you've added)* | Dynamic column sourced from an agent's input parameters or structured output |

<Info>
  The **LLM**, **LLM detail**, **Avg AI latency**, **Max latency**, and **Latency (turns)** columns are optional. In the classic Call History table they're hidden by default and can be enabled from the column customizer; in the improved calls view, **Avg AI latency** and **Latency (turns)** are visible out of the box.
</Info>

### LLM and LLM Detail Columns

<Info>
  These two columns surface the underlying language model activity for every call, which is especially useful for agents built on ElevenLabs' conversational stack, where a **model cascade** can automatically fall back to a secondary model mid-call if the primary one becomes unavailable or degrades.
</Info>

| Column         | What it shows                                                                                                                                                                                    |
| -------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| **LLM**        | The distinct model(s) used during the call, listed in the order they were engaged (e.g. a single model like `gpt-4o`, or multiple models such as `gpt-4o + gpt-4o-mini` if a fallback occurred). |
| **LLM detail** | An expandable, turn-by-turn breakdown showing exactly which model handled each conversational turn and at what point in the call the switch happened.                                            |

<Steps>
  <Step title="Enable the Columns">
    Open the column manager above the Call Table and toggle on **LLM** and/or **LLM detail** to add them to your view.
  </Step>

  <Step title="Read the LLM Column">
    Each row shows the model(s) used for that call. A single model name means the call ran entirely on the primary model with no fallback.
  </Step>

  <Step title="Spot Degradations">
    When a call shows **more than one model** in the LLM column, it means the ElevenLabs cascade degraded at least once during the call — falling back from the primary model to a secondary one. This is the same detection logic used in the call details modal.
  </Step>

  <Step title="Drill Into LLM Detail">
    Expand the LLM detail column (or open the call in the Details Modal) to see the exact timestamp of each turn and which model produced it, helping you pinpoint when and why a degradation happened.
  </Step>
</Steps>

<Tip>
  Use the LLM column together with the Status and Success filters to quickly check whether cascade degradations correlate with lower success rates or specific agents, then dig deeper with LLM detail on the affected calls.
</Tip>

<Warning>
  For calls made before model cascade tracking was introduced, the LLM column falls back to the legacy single `model_used` value when no multi-model data is available. In that case, no degradation can be detected for those older calls.
</Warning>

### Bulk Actions

<Tabs>
  <Tab title="Selection">
    When you select multiple calls:

    1. The **"Export selection (N)"** button appears in the header, next to the general **Export to Excel** button
    2. Bulk export options become available immediately, no separate confirmation step required
    3. Additional bulk actions appear in the action bar
    4. Deselect individual rows or clear the whole selection to hide the selective export button again
  </Tab>

  <Tab title="Export">
    Export filtered results:

    * Excel export that follows your currently visible columns, their order, and review status
    * CSV format for spreadsheet analysis (classic view)
    * JSON format for technical analysis (classic view)
    * PDF reports for presentations (classic view)
    * Include/exclude specific columns via the column manager before exporting
  </Tab>
</Tabs>

### Enhanced Sorting

Click column headers to sort with improved indicators:

* **Single Click** - Ascending order with ↑ arrow
* **Double Click** - Descending order with ↓ arrow
* **Sort Memory** - Last sort preference is remembered
* **Multi-Column Sort** - Hold Shift to sort by multiple columns

## Pagination

Enhanced pagination controls:

* **Smart Page Sizes** - 25, 50, 100, 300 calls per page
* **Jump to Page** - Direct page number input
* **Results Counter** - "Showing X to Y of Z calls" with total count
* **Performance Optimization** - Efficient loading for large datasets

## Quick Actions

For each call row with improved accessibility:

| Action            | Description                        | Shortcut                         |
| ----------------- | ---------------------------------- | -------------------------------- |
| 👁️ View          | Open call details modal            | Click row                        |
| 📋 Copy ID        | Copy call ID to clipboard          | Copy button on the Call ID badge |
| 🏷️ Add Tags      | Add custom tags to the call        | Tag icon                         |
| 📊 Analytics      | Quick performance insights         | Chart icon                       |
| ▶️ Play Recording | Open the waveform recording player | Recording column                 |
|                   |                                    |                                  |


# Reporting System
Source: https://docs.meetzy.es/calls/reporting

Generate advanced reports with agent metrics, temporal analysis, and export capabilities

The Reporting System provides comprehensive analytics for your conversational AI agents with advanced filtering, metrics tracking, and data export capabilities.

## Overview

The reporting system offers detailed insights into agent performance, call metrics, and conversion analytics with temporal analysis to help you optimize your AI agents.

<Info>
  All reports are generated in real-time and can be exported to CSV format for further analysis.
</Info>

## Getting Started

<Steps>
  <Step title="Access Reports">
    Navigate to the Reports section in your dashboard to begin generating analytics reports.
  </Step>

  <Step title="Select Agents">
    Choose which agents you want to include in your report from the available list.
  </Step>

  <Step title="Configure Filters">
    Set date ranges, status filters, and other criteria to focus your analysis.
  </Step>

  <Step title="Generate Report">
    Click "Generate Report" to create your comprehensive analytics report.
  </Step>
</Steps>

## Report Filters

### Basic Filters

Configure your report scope with these fundamental filters:

* **Agent Selection**: Choose specific agents or select all agents
* **Date Range**: Set start and end dates for your analysis period
* **Status Filter**: Filter by call status (completed, missed, failed, etc.)
* **Call Type**: Filter by inbound, outbound, or transferred calls

### Advanced Filters

Refine your analysis with additional criteria:

```typescript theme={null}
interface ReportFilters {
  agent_ids: string[];
  start_date: string;
  end_date: string;
  filters?: {
    status?: string[];
    call_type?: string[];
    min_duration?: number;
    max_duration?: number;
    min_nps_score?: number;
    max_nps_score?: number;
  };
}
```

**Duration Filters**:

* Set minimum and maximum call duration thresholds
* Analyze calls within specific time ranges

**NPS Score Filters**:

* Filter by Net Promoter Score ranges
* Focus on high or low satisfaction calls

## Agent Metrics

Each report provides comprehensive metrics for selected agents:

### Core Metrics

<Tabs>
  <Tab title="Call Statistics">
    * **Total Calls**: All initiated calls
    * **Completed Calls**: Successfully finished conversations
    * **Converted Calls**: Calls that achieved objectives
    * **Received Calls**: Inbound calls handled
    * **Outbound Calls**: Agent-initiated calls
    * **Transferred Calls**: Calls passed to human agents
  </Tab>

  <Tab title="Performance Metrics">
    * **Completion Rate**: Percentage of calls completed successfully
    * **Conversion Rate**: Percentage of calls that converted
    * **Contact Rate**: Percentage of leads successfully contacted
    * **Transfer Rate**: Percentage of calls transferred to humans
  </Tab>

  <Tab title="Duration Analysis">
    * **Average Duration**: Mean call length
    * **Total Conversation Time**: Sum of all call durations
    * **Duration Standard Deviation**: Variability in call lengths
    * **Billable Minutes**: Total charged conversation time
  </Tab>
</Tabs>

### AI Response Latency

<Info>
  Response latency measures how long the AI agent takes to start speaking after the caller stops talking. Lower latency generally means a more natural, human-like conversation.
</Info>

In the calls table view, two latency metrics are available as **columns** for every call:

* **Avg AI latency**: The average response latency (in milliseconds) across all agent turns in that call.
* **Max latency**: The highest single response latency (in milliseconds) recorded during that call.

A compact **Latency (turns)** sparkline column is also available, giving you a quick visual trend of how latency evolved turn-by-turn throughout the call. Clicking into a call's details opens a full latency chart with the exact value for each turn.

<Info>
  Both metrics can also be used as **filters** in the calls view (`Min latency (ms)` / `Max latency (ms)`) to quickly surface calls where the AI took unusually long to respond — useful for troubleshooting perceived sluggishness or validating performance after an infrastructure change.
</Info>

By default, **Avg AI latency** and **Latency (turns)** are shown as visible columns, while **Max latency** is available but hidden — you can enable it anytime from the column manager (⚙️ Columns).

### Customer Satisfaction

<Info>
  NPS (Net Promoter Score) tracking is available for agents with NPS activation enabled.
</Info>

* **Average NPS Score**: Mean customer satisfaction rating
* **NPS Surveys Count**: Number of completed satisfaction surveys
* **NPS Response Rate**: Percentage of customers who responded to surveys

### Lead Management

Track your lead conversion funnel:

* **Unique Leads**: Total number of distinct leads
* **Contacted Leads**: Leads successfully reached
* **Converted Leads**: Leads that completed desired actions
* **Conversion from Contacted Rate**: Efficiency of contacted lead conversion

## Temporal Analysis

### Daily Breakdown

View performance trends over time with daily metrics:

```typescript theme={null}
daily_data: Array<{
  date: string;
  calls: number;
  unique_leads: number;
  contacted_leads: number;
  converted_leads: number;
  completed_calls: number;
  conversions: number;
  avg_duration: number;
  conversion_rate: number;
  // Additional metrics...
}>
```

**Key Daily Metrics**:

* Call volume trends
* Conversion rate fluctuations
* Duration patterns
* Lead engagement levels

### Hourly Analysis

Optimize scheduling with hourly performance data:

* **Peak Hours**: Identify most effective calling times
* **Conversion by Hour**: Success rates throughout the day
* **Duration Patterns**: Average call lengths by time of day

<Warning>
  Hourly analysis requires sufficient data volume for meaningful insights. Reports with fewer than 50 calls may show limited hourly patterns.
</Warning>

## Data Export

### CSV Export

Export your reports for external analysis:

<Steps>
  <Step title="Generate Report">
    Create your report with desired filters and date range.
  </Step>

  <Step title="Export Options">
    Choose CSV format and optionally specify a custom filename.
  </Step>

  <Step title="Download">
    The system will generate and download your CSV file with all report data.
  </Step>
</Steps>

**CSV Export Includes**:

* Agent metrics summary
* Daily temporal data
* Hourly analysis
* Status distribution
* Call details (if included in filters)

### Export Format

```typescript theme={null}
// Export request structure
{
  report_data: ReportData,
  format: 'csv',
  filename?: string // Optional custom filename
}
```

## Report Templates

### Predefined Templates

Use pre-configured templates for common reporting needs:

* **Daily Performance**: Standard daily metrics overview
* **Weekly Summary**: 7-day performance analysis
* **Monthly Review**: Comprehensive monthly report
* **Conversion Analysis**: Focus on conversion metrics
* **NPS Analysis**: Customer satisfaction report

### Custom Templates

Create and save your own report configurations:

```typescript theme={null}
interface ReportTemplate {
  id: string;
  name: string;
  description: string;
  filters: ReportFilters;
}
```

## Status Distribution

Understand call outcomes with status distribution analysis:

* **Completion Status**: Breakdown of finished, ongoing, and failed calls
* **Contact Results**: Distribution of answered, missed, and unreachable calls
* **Percentage Analysis**: Visual representation of status proportions

Common statuses include:

* `closed`: Successfully finished calls
* `open`: Currently ongoing calls
* `no-answer`: Unanswered calls
* `voicemail`: Calls reaching voicemail
* `failed`: Technical failures
* `busy`: Busy line encounters

## Best Practices

### Effective Reporting

<Info>
  For optimal insights, use date ranges that provide sufficient data while remaining focused on your analysis goals.
</Info>

1. **Regular Monitoring**: Generate weekly reports to track performance trends
2. **Comparative Analysis**: Compare different time periods to identify improvements
3. **Agent Benchmarking**: Compare agent performance to identify best practices
4. **Seasonal Adjustments**: Account for seasonal variations in your analysis

### Data Interpretation

* **Look for Patterns**: Identify recurring trends in daily and hourly data
* **Context Matters**: Consider external factors affecting call performance
* **Actionable Insights**: Focus on metrics you can improve through optimization

### Performance Optimization

Use reporting data to:

* Adjust calling schedules based on hourly success rates
* Optimize agent scripts based on conversion analysis
* Improve lead quality based on contact rate insights
* Enhance customer experience using NPS feedback
* Spot outlier calls with high AI response latency and investigate potential causes (network conditions, model load, prompt complexity)

## API Integration

The Reporting Service provides programmatic access to all reporting functionality:

```typescript theme={null}
// Generate a report
const reportData = await reportingService.generateReport(filters);

// Export report data
const csvBlob = await reportingService.exportReport(reportData, 'csv');

// Get available agents
const agents = await reportingService.getAgents();

// Access report templates
const templates = await reportingService.getTemplates();
```

<Warning>
  API access requires proper authentication tokens. Ensure your integration handles token refresh for long-running processes.
</Warning>


# Campaigns
Source: https://docs.meetzy.es/crm/campaigns

Schedule and manage mass calling campaigns to contact segments

# Campaigns

Campaigns allow you to schedule mass outbound calls to segments of contacts. Your AI agents will automatically call each contact in the segment based on your configuration.

## What is a Campaign?

A campaign connects:

| Component   | Description                           |
| ----------- | ------------------------------------- |
| 🤖 Agent    | The AI agent that will make the calls |
| 👥 Segment  | The group of contacts to call         |
| ⏰ Schedule  | When calls should be made             |
| ⚙️ Settings | Call limits, retries, and more        |

## Creating a Campaign

<Steps>
  <Step title="Select Agent">
    Choose which AI agent will handle the calls.
  </Step>

  <Step title="Choose Segment">
    Select the contact segment to target.
  </Step>

  <Step title="Configure Schedule">
    Set when calls should be made.
  </Step>

  <Step title="Set Options">
    Configure call limits and retry rules.
  </Step>

  <Step title="Launch">
    Start or schedule the campaign.
  </Step>
</Steps>

## Campaign Settings

### Agent Selection

* Choose from your active agents
* Agent must have an assigned phone number
* Outbound greeting is used for these calls

### Segment Selection

* Select from available segments
* View contact count before launching
* Contacts without phone numbers are skipped

### Schedule Options

| Option            | Description                    |
| ----------------- | ------------------------------ |
| Start Immediately | Begin calling right away       |
| Schedule Start    | Set a specific start date/time |
| Calling Hours     | Limit calls to specific hours  |
| Time Zone         | Define which time zone to use  |

<Tip>
  Set appropriate calling hours based on your contacts' time zones. Calling during business hours improves answer rates.
</Tip>

### Call Settings

| Setting          | Description                        |
| ---------------- | ---------------------------------- |
| Calls per Hour   | Maximum calls to make per hour     |
| Max Attempts     | How many times to try each contact |
| Retry Delay      | Time between retry attempts        |
| Concurrent Calls | Maximum simultaneous calls         |

## Campaign Status

| Status       | Description                         |
| ------------ | ----------------------------------- |
| 🟡 Draft     | Campaign configured but not started |
| 🟢 Active    | Campaign is currently running       |
| ⏸️ Paused    | Campaign temporarily stopped        |
| ✅ Completed  | All contacts have been called       |
| 🔴 Cancelled | Campaign was stopped                |

## Managing Campaigns

### Starting a Campaign

1. Review all settings
2. Click **Launch Campaign**
3. Confirm the start
4. Campaign begins processing contacts

### Pausing a Campaign

1. Click **Pause** on an active campaign
2. Active calls complete, new calls stop
3. Resume anytime with **Resume**

### Cancelling a Campaign

1. Click **Cancel** or **Stop**
2. Confirm the cancellation
3. Campaign cannot be restarted (create a new one)

## Campaign Progress

Monitor your campaign in real-time:

| Metric         | Description                   |
| -------------- | ----------------------------- |
| 📊 Total       | Total contacts in segment     |
| ✅ Completed    | Successfully called contacts  |
| ⏳ Pending      | Contacts waiting to be called |
| ❌ Failed       | Calls that couldn't complete  |
| 🔄 In Progress | Currently active calls        |

## Call Results

For each contact, track:

| Result    | Description                              |
| --------- | ---------------------------------------- |
| Answered  | Contact picked up, conversation occurred |
| No Answer | No one answered the call                 |
| Busy      | Line was busy                            |
| Voicemail | Call went to voicemail                   |
| Failed    | Technical error prevented the call       |
| Completed | Call finished with a result              |

## Retry Logic

### Automatic Retries

When a call fails or goes unanswered:

1. Wait for the configured retry delay
2. Attempt the call again
3. Continue until max attempts reached
4. Mark as failed after all attempts exhausted

### Retry Settings

| Setting      | Recommendation       |
| ------------ | -------------------- |
| Max Attempts | 2-3 attempts         |
| Retry Delay  | 30 minutes - 2 hours |
| Time of Day  | Vary retry times     |

## Campaign Analytics

After a campaign completes:

| Metric           | Description                          |
| ---------------- | ------------------------------------ |
| Answer Rate      | Percentage of calls answered         |
| Conversion Rate  | Successful outcomes / Total answered |
| Average Duration | Mean call length                     |
| Best Times       | Hours with highest answer rates      |

## Best Practices

<CardGroup>
  <Card title="Test First" icon="flask">
    Run a small test campaign before launching to thousands.
  </Card>

  <Card title="Respect Hours" icon="clock">
    Only call during appropriate business hours.
  </Card>

  <Card title="Monitor Progress" icon="chart-line">
    Check campaign progress regularly for issues.
  </Card>

  <Card title="Segment Well" icon="layer-group">
    Smaller, targeted segments perform better.
  </Card>
</CardGroup>

## Compliance Considerations

<Warning>
  Ensure your campaigns comply with local regulations (TCPA in US, GDPR in EU). Only call contacts who have consented to receive calls.
</Warning>

### Compliance Checklist

* ✅ Contacts have opt-in consent
* ✅ Calling during permitted hours
* ✅ Providing opt-out option
* ✅ Honoring do-not-call requests
* ✅ Keeping consent records

## Troubleshooting

| Issue                | Solution                                         |
| -------------------- | ------------------------------------------------ |
| Low answer rate      | Adjust calling hours, check caller ID reputation |
| High failure rate    | Verify phone numbers, check agent phone health   |
| Slow progress        | Increase concurrent calls or calls per hour      |
| Campaign won't start | Ensure agent has active phone number             |

## Next Steps

<CardGroup>
  <Card title="Segments" icon="layer-group" href="/crm/segments">
    Create targeted segments for your campaigns.
  </Card>

  <Card title="Phone Health" icon="heart-pulse" href="/agents/phone-health">
    Monitor your phone number reputation.
  </Card>
</CardGroup>


# Import Contacts
Source: https://docs.meetzy.es/crm/import-contacts

Import contacts from Excel files or Google Sheets with advanced filters and AI-powered search

# Import Contacts

Meetzy allows you to import contacts in bulk from Excel files (.xlsx, .xls, .xlsm) or directly from Google Sheets. This enables you to quickly populate your contact database for campaigns with advanced filtering and AI-powered search capabilities.

## Import Methods

| Method           | Description                           | Requirements               |
| ---------------- | ------------------------------------- | -------------------------- |
| 📊 Excel         | Upload .xlsx, .xls, or .xlsm files    | File up to 10MB            |
| 📗 Google Sheets | Connect directly to your spreadsheets | Google account integration |

## Import Wizard

The import process follows three steps:

<Steps>
  <Step title="Upload / Select">
    Choose your source file or spreadsheet.
  </Step>

  <Step title="Configure">
    Map columns and set import options with advanced filters.
  </Step>

  <Step title="Import">
    Process contacts and view results.
  </Step>
</Steps>

## Excel Import

### Uploading a File

1. Navigate to **CRM** → **Import Contacts**
2. Select **Import from Excel**
3. Drag and drop your file or click to browse
4. Supported formats: `.xlsx`, `.xls`, `.xlsm`
5. Maximum file size: 10MB

<Info>
  Drag and drop works directly on the upload zone. The border turns green when a valid file is detected.
</Info>

### File Validation

Once uploaded, the system analyzes your file:

| Analysis     | Description                   |
| ------------ | ----------------------------- |
| Total Rows   | Number of data rows detected  |
| Valid Phones | Rows with valid phone numbers |
| Duplicates   | Duplicate phone numbers found |
| No Phone     | Rows missing phone numbers    |

<Warning>
  Rows without phone numbers will be skipped during import. Ensure your phone column is properly formatted.
</Warning>

## Google Sheets Import

### Connecting Google Sheets

If not connected:

1. Click **Connect Google Sheets**
2. Authorize with your Google account
3. Grant access to your spreadsheets

### Selecting a Spreadsheet

1. Browse your available spreadsheets
2. Use the search bar to find specific files
3. Click on a spreadsheet to select it
4. The system validates the sheet structure

## Advanced Filters & Search

### AI-Powered Search

Use natural language to find specific contacts:

<Steps>
  <Step title="Open AI Search">
    Click the AI search icon in the search bar.
  </Step>

  <Step title="Enter Query">
    Type your search in natural language, e.g., "Find all contacts from California with more than 5 calls"
  </Step>

  <Step title="Review Results">
    AI will process your query and display matching contacts.
  </Step>
</Steps>

<Info>
  AI search works with contact data, call history, location information, and custom fields to provide intelligent results.
</Info>

### Advanced Filtering

Access advanced filters to narrow down your import data:

| Filter Type      | Options               | Description                             |
| ---------------- | --------------------- | --------------------------------------- |
| **Status**       | Active, Inactive, All | Filter by contact status                |
| **Call Range**   | Min/Max calls         | Filter by number of previous calls      |
| **Date Range**   | From/To dates         | Filter by creation or last contact date |
| **Location**     | City, Country         | Filter by geographic location           |
| **Contact Info** | Has Email             | Filter contacts with/without email      |

### Smart Filter Combinations

Combine multiple filters for precise targeting:

```
Status: Active + 
Min Calls: 3 + 
Has Email: Yes + 
City: Madrid
```

This finds active contacts from Madrid with at least 3 calls and email addresses.

## Column Mapping

### Required Fields

| Field | Description          | Required |
| ----- | -------------------- | -------- |
| Phone | Contact phone number | ✅ Yes    |
| Name  | Contact name         | No       |
| Email | Contact email        | No       |

### Enhanced Mapping Options

1. **Standard Fields**: Phone, Name, Email
2. **Location Fields**: City, State, Country, Postal Code
3. **Business Fields**: Company, Job Title
4. **Marketing Fields**: UTM Source, Medium, Campaign
5. **Custom Fields**: Any additional columns from your file

### Smart Column Detection

The system automatically suggests column mappings based on:

* Column headers
* Data patterns
* Common naming conventions

<Info>
  The system recognizes headers like "First Name", "fname", "given\_name" and maps them appropriately.
</Info>

### Custom Fields

Add additional columns as custom fields:

1. View available unmapped columns
2. Click on a column name to include it
3. Custom fields are stored with the contact
4. Use them in agent prompts with input parameters

<Warning>
  Columns already mapped to standard fields (Phone, Name, Email) cannot be added as custom fields.
</Warning>

## Data Preview & Validation

### Enhanced Preview

Before importing, review your data with advanced preview features:

* **Data Quality Indicators**: Shows data completeness and quality scores
* **Duplicate Detection**: Highlights potential duplicate contacts
* **Format Validation**: Verifies phone numbers and email formats
* **Geographic Validation**: Validates city/country combinations

### Data Quality Metrics

| Metric     | Description                | Action                 |
| ---------- | -------------------------- | ---------------------- |
| ✅ Valid    | Properly formatted data    | Ready to import        |
| ⚠️ Warning | Potentially incorrect data | Review recommended     |
| ❌ Error    | Invalid data format        | Must fix before import |

## AI-Enhanced Data Enrichment

### Automatic Data Enhancement

During import, the system can automatically:

1. **Format phone numbers** with proper country codes
2. **Standardize addresses** using geographic databases
3. **Detect duplicates** using fuzzy matching algorithms
4. **Enrich contact data** with additional information

### Intelligent Duplicate Handling

Choose how to handle duplicates:

| Option     | Description              | Recommendation        |
| ---------- | ------------------------ | --------------------- |
| Skip       | Skip duplicate entries   | Best for clean data   |
| Update     | Update existing contacts | Good for data refresh |
| Create New | Create duplicate entries | Use with caution      |

## Segment Assignment

All imported contacts are automatically added to a segment:

1. Enter a **Segment Name**
2. This creates a new segment with your imports
3. Use segments for targeted campaigns
4. Apply advanced filters to segment creation

<Info>
  Use descriptive segment names like "Q1 2024 Leads - Madrid Tech Companies" for easy identification and targeting.
</Info>

### Smart Segment Suggestions

Based on your data, the system suggests segment names:

* **By Location**: "Madrid Contacts", "Spain Prospects"
* **By Source**: "LinkedIn Import March 2024"
* **By Industry**: "Tech Companies", "Real Estate Leads"

## Admin Features

For admin users:

### Company Selection

1. A company selector appears at the top
2. Search for a company by name with intelligent autocomplete
3. Select to import contacts to that company
4. "My Account" imports to your own account

### Bulk Import Analytics

View detailed analytics during import:

* **Import velocity** (contacts per minute)
* **Data quality scores**
* **Enrichment statistics**
* **Duplicate resolution summary**

## Import Process

Once configuration is complete:

1. Click **Start Import**
2. Real-time progress with advanced metrics
3. Live stats display:
   * New contacts created
   * Existing contacts updated
   * Data enrichment applied
   * Rows skipped with reasons

### Enhanced Import Duration

Processing includes:

* **Data validation** and formatting
* **Duplicate detection** and resolution
* **AI-powered data enrichment**
* **Advanced segmentation**

Large imports (10,000+ rows) may take several minutes depending on enrichment settings.

## Import Report

After completion, you'll see an enhanced report:

| Metric        | Description                             |
| ------------- | --------------------------------------- |
| ✅ Imported    | New contacts created successfully       |
| 🔄 Updated    | Existing contacts updated with new data |
| 🤖 Enriched   | Contacts enhanced with additional data  |
| ⏭️ Skipped    | Rows not imported with detailed reasons |
| 📊 In Segment | Total contacts added to segment         |

### Advanced Error Analysis

If errors occurred:

* **Categorized error types** (format, validation, duplicate)
* **Suggested fixes** for common issues
* **Bulk correction options** for similar errors
* **Export error report** for offline review

### Post-Import Actions

* **View Segment** - Go to the created segment with applied filters
* **Run AI Analysis** - Get insights about imported contacts
* **Start Campaign** - Create targeted campaigns immediately
* **Export Results** - Download import summary report

## AI-Powered Insights

After import, access AI insights about your new contacts:

<CardGroup>
  <Card title="Contact Analysis" icon="chart-line">
    Geographic distribution, industry breakdown, and quality metrics.
  </Card>

  <Card title="Engagement Prediction" icon="crystal-ball">
    AI predicts which contacts are most likely to engage.
  </Card>

  <Card title="Segmentation Suggestions" icon="layer-group">
    Smart recommendations for additional segment creation.
  </Card>

  <Card title="Campaign Optimization" icon="target">
    Optimal calling times and approach strategies.
  </Card>
</CardGroup>

## Best Practices

<CardGroup>
  <Card title="Data Quality First" icon="shield-check">
    Use advanced filters to ensure high-quality imports with proper validation.
  </Card>

  <Card title="Smart Segmentation" icon="layer-group">
    Leverage AI suggestions for creating meaningful contact segments.
  </Card>

  <Card title="Duplicate Prevention" icon="copy">
    Enable intelligent duplicate detection to maintain clean databases.
  </Card>

  <Card title="Gradual Enrichment" icon="sparkles">
    Start with basic imports, then use AI enhancement for progressive data improvement.
  </Card>
</CardGroup>

## Troubleshooting

| Issue                     | Advanced Solution                                      |
| ------------------------- | ------------------------------------------------------ |
| "AI search not working"   | Check internet connection and try simpler queries      |
| "Advanced filters slow"   | Reduce filter complexity or import in smaller batches  |
| "Data enrichment failed"  | Verify data quality and contact support for API limits |
| "Smart duplicates missed" | Adjust duplicate detection sensitivity in settings     |

## Next Steps

<CardGroup>
  <Card title="AI Contact Search" icon="magnifying-glass" href="/crm/ai-search">
    Use AI to find specific contacts with natural language queries.
  </Card>

  <Card title="Advanced Segments" icon="layer-group" href="/crm/segments">
    Create intelligent segments using advanced filters and AI.
  </Card>

  <Card title="Campaign Optimization" icon="rocket" href="/campaigns/optimization">
    Use AI insights to optimize your calling campaigns.
  </Card>

  <Card title="Data Analytics" icon="chart-bar" href="/analytics/contacts">
    Analyze contact data with advanced reporting tools.
  </Card>
</CardGroup>


# Segments
Source: https://docs.meetzy.es/crm/segments

Create and manage contact groups for targeted campaigns

# Segments

Segments are groups of contacts that you can use for targeted campaigns. Organize your contacts by criteria like lead source, geography, or customer type.

## Accessing Segments

Navigate to **CRM** → **Segments** to view all your contact segments.

## Segments Overview

The segments page displays:

| Column      | Description                   |
| ----------- | ----------------------------- |
| Name        | Segment identifier            |
| Description | Optional description          |
| Contacts    | Number of contacts in segment |
| Created     | When the segment was created  |
| Actions     | View, Edit, Delete options    |

## Creating a Segment

<Steps>
  <Step title="Click Create Segment">
    Click the **Create Segment** button in the top right.
  </Step>

  <Step title="Enter Details">
    * **Segment Name** - Required, identifies the segment
    * **Description** - Optional, explain the segment's purpose
  </Step>

  <Step title="Save">
    Click **Create** to save your segment.
  </Step>
</Steps>

<Tip>
  Create segments before importing contacts, or contacts will be added to an auto-generated segment.
</Tip>

## Segment Properties

| Property      | Description                                |
| ------------- | ------------------------------------------ |
| Name          | Unique identifier for the segment          |
| Description   | Explains what contacts are in this segment |
| Contact Count | Number of contacts (updated automatically) |
| Created Time  | When the segment was first created         |

## Managing Segments

### Viewing Contacts

1. Click **View** on any segment row
2. Opens the segment detail view
3. See all contacts within the segment
4. Search and filter contacts

### Editing a Segment

1. Click the **Edit** (pencil) icon
2. Modify name or description
3. Click **Save Changes**

<Info>
  Editing a segment's name or description doesn't affect the contacts within it.
</Info>

### Deleting a Segment

1. Click the **Delete** (trash) icon
2. Confirm the deletion
3. Contacts are NOT deleted, only removed from the segment

<Warning>
  Segment deletion cannot be undone. Make sure no active campaigns depend on this segment.
</Warning>

## Searching Segments

Use the search bar to find segments:

* Search by segment name
* Search by description
* Results filter in real-time

## Sorting Segments

Click column headers to sort:

| Column  | Sort Options               |
| ------- | -------------------------- |
| Name    | A-Z, Z-A                   |
| Created | Newest first, Oldest first |

## Admin Features

For admin users:

### Company Filter

1. A company filter appears at the top
2. Search for companies
3. Select a company to view their segments
4. "All Companies" shows segments across all

### Cross-Company Access

Admins can:

* View segments from any company
* Create segments for specific companies
* Manage campaigns across accounts

## Using Segments

### In Campaigns

Segments are used to target campaigns:

1. Create a campaign
2. Select a segment as the target audience
3. All contacts in that segment receive calls

### Segment Size

For optimal campaign performance:

| Size         | Recommendation                  |
| ------------ | ------------------------------- |
| \< 100       | Good for testing                |
| 100-1,000    | Standard campaign size          |
| 1,000-10,000 | Large campaign, monitor closely |
| > 10,000     | Split into multiple campaigns   |

## Segment Strategies

<CardGroup>
  <Card title="By Source" icon="door-open">
    Group contacts by how they were acquired (web form, import, API).
  </Card>

  <Card title="By Region" icon="map">
    Organize by geography for localized campaigns.
  </Card>

  <Card title="By Status" icon="circle-check">
    Separate qualified leads from cold contacts.
  </Card>

  <Card title="By Date" icon="calendar">
    Group by import date for time-based follow-ups.
  </Card>
</CardGroup>

## Best Practices

1. **Descriptive Names** - Use clear names like "January 2024 Webinar Signups"
2. **Add Descriptions** - Document criteria for future reference
3. **Regular Cleanup** - Archive or delete old segments
4. **Size Monitoring** - Keep segment sizes manageable for campaigns

## Integration with Imports

When importing contacts:

1. Specify a segment name during import
2. All contacts are added to that segment
3. Or select an existing segment to append to

## Next Steps

<CardGroup>
  <Card title="Import Contacts" icon="file-import" href="/crm/import-contacts">
    Add contacts to your segments.
  </Card>

  <Card title="Campaigns" icon="bullhorn" href="/crm/campaigns">
    Use segments to launch targeted call campaigns.
  </Card>
</CardGroup>


# Create Your First Agent
Source: https://docs.meetzy.es/getting-started/create-agent

Complete guide to creating and configuring an AI voice agent in Meetzy

## Creating an Agent

Meetzy offers two ways to create your AI voice agent:

<Tabs>
  <Tab title="Prompt-First Creation (Recommended)">
    The fastest way to get started. Simply describe what your agent should do, and Meetzy builds it for you automatically.
  </Tab>

  <Tab title="Advanced Editor">
    Full control over every aspect of your agent's configuration using our three-panel interface.
  </Tab>
</Tabs>

## Prompt-First Creation

As soon as you click **"New Agent"**, Meetzy no longer drops you directly into an empty editor. Instead, a full-screen creation overlay appears — a welcoming screen with an animated wave background — where you describe your agent in natural language and let Meetzy Copilot generate the initial configuration for you.

<Info>
  Describe what you want, press Enter, and Meetzy builds a working agent automatically. This is the recommended starting point for both new and experienced users.
</Info>

### The Creation Screen

The overlay is intentionally minimal and distraction-free:

* A centered heading, **"What should your agent do?"**, with the subtitle *"Describe it and let Meetzy build it for you."*
* A large textarea with a soft gradient border, auto-focused so you can start typing immediately.
* A send button (arrow icon) that fades in as soon as you type something.
* A subtle animated wave background rendered on canvas, which automatically adapts to your light/dark theme.
* The Meetzy logo centered at the top, and a back button in the top-left corner to return to the agents list.

<Steps>
  <Step title="Describe Your Agent">
    In the text area, describe what you want your agent to do. Be specific about:

    * **Purpose**: What is the agent's main function?
    * **Industry/Context**: What business or service does it support?
    * **Key Actions**: What should it be able to do?
    * **Escalation**: When should it transfer to humans?

    Example descriptions:

    ```
    An agent that answers customer support calls for a dental clinic, 
    books appointments, and transfers urgent cases to the receptionist.
    ```

    ```
    A sales qualification bot for a real estate company that qualifies 
    leads, schedules property viewings, and collects contact information.
    ```

    ```
    Customer service agent for an e-commerce store that handles order 
    inquiries, processes returns, and provides shipping updates.
    ```
  </Step>

  <Step title="Generate Your Agent">
    Press **Enter** or click the send button (arrow icon) to launch the generation. Meetzy Copilot will automatically:

    * Generate an appropriate system prompt
    * Configure basic settings and phone integration
    * Set up default greetings for inbound/outbound calls
    * Create relevant input/output parameters
    * Add suggested tools and actions

    While your description is processed, the overlay animates out and the Advanced Editor fades in with your new configuration already applied.
  </Step>

  <Step title="Review and Customize">
    Your agent opens in the Advanced Editor where you can:

    * Review the generated configuration
    * Customize the system prompt
    * Adjust settings and parameters
    * Test with the playground
    * Publish when ready
  </Step>
</Steps>

### Skipping the Prompt-First Flow

If you'd rather build your agent manually, section by section, you don't have to write a description at all.

<Info>
  Click **"or start from scratch"** below the textarea (or press **Escape**) to dismiss the creation overlay instantly and land directly in the Advanced Editor with a blank configuration.
</Info>

This shortcut is ideal for advanced users who already know exactly what they want to configure, or when you're duplicating patterns you've used before. The back arrow in the top-left corner also skips the overlay and returns you to the agents list.

### Creation Screen Features

* **Animated Background**: A lightweight canvas-based wave animation that adapts to your theme
* **Keyboard Shortcuts**:
  * Press `Enter` to generate your agent from the description
  * Press `Escape` to skip straight to the Advanced Editor
* **Theme Aware**: Automatically matches your light/dark theme preference
* **Mobile Friendly**: Works perfectly on mobile devices
* **Auto-focus**: The textarea is focused automatically so you can start typing right away
* **Smooth Transition**: The overlay animates out gracefully once you submit your prompt, revealing the editor underneath

<Warning>
  The creation overlay only appears when starting a brand-new agent from scratch. If you're editing an existing agent, or if your configuration was already pre-populated (for example, through onboarding), you'll be taken directly to the Advanced Editor.
</Warning>

## Automatic Agent Creation from Onboarding

When you complete the onboarding flow from our self-service landing page, Meetzy automatically creates your first agent to streamline your setup process.

<Info>
  If you came from our onboarding process, your agent configuration has been automatically preserved and you'll be redirected directly to the agent editor with your initial settings applied — the prompt-first creation overlay is skipped in this case, since your description was already captured during onboarding.
</Info>

### How It Works

<Steps>
  <Step title="Onboarding Data Preservation">
    The system automatically saves your onboarding configuration including:

    * Initial prompt description
    * Selected communication channels
    * Business directions and requirements
  </Step>

  <Step title="Automatic Redirection">
    After successful authentication, you're automatically redirected to the agent creation editor instead of the standard dashboard
  </Step>

  <Step title="Pre-populated Configuration">
    Your agent editor opens with your onboarding data already applied, so you can immediately review and customize your agent
  </Step>
</Steps>

<Warning>
  Onboarding data is automatically cleared after use to maintain security. If you navigate away before completing your agent setup, you'll need to manually re-enter your configuration.
</Warning>

## Prerequisites

Before creating an agent, ensure you have:

* An active Meetzy account with dashboard access
* At least one phone number available in your account
* Basic understanding of your use case (sales, support, etc.)

## The Advanced Editor

The Advanced Editor is a three-panel interface for configuring your agent:

| Panel              | Description                                      |
| ------------------ | ------------------------------------------------ |
| **Left Panel**     | Navigation between configuration sections        |
| **Central Editor** | Main editing area for system prompt and settings |
| **Right Panel**    | Playground for testing and debugging             |

### Toolbar Actions

| Button                | Description                      |
| --------------------- | -------------------------------- |
| ← **Back**            | Return to agents list            |
| 🏷️ **Version Badge** | Current version (e.g., v.1, v.2) |
| ✓ **Check**           | Validate configuration           |
| ⬇️ **Pull**           | Sync with latest version         |
| 🚀 **Publish**        | Deploy to production             |

## Left Panel

The left sidebar contains all configuration sections:

| Icon | Section             | Description                                |
| ---- | ------------------- | ------------------------------------------ |
| ⚙️   | **Settings**        | Basic agent configuration                  |
| 👋   | **Greetings**       | Welcome messages for calls                 |
| 📄   | **Input Params**    | Dynamic variables for personalization      |
| 📚   | **Knowledge Base**  | Documents and FAQs the agent can reference |
| 🔧   | **Actions & Tools** | Integrations and external APIs             |
| ✅    | **Evaluations**     | Automated tests to validate agent behavior |
| 📥   | **Output Settings** | Data the agent should capture              |

## Step 1: Basic Settings

Click the **Settings** icon (Cog) to configure:

### Agent Name

Internal identifier for your agent. Choose a descriptive name like:

* "Sales Qualification Agent"
* "Customer Support - Billing"
* "Appointment Reminder"

### Company

Select which company this agent belongs to (for multi-company accounts).

### Twilio Account

Choose the Twilio account for phone calls.

### Phone Number

<Warning>
  Phone numbers are exclusive resources. Each number can only be assigned to one agent at a time.
</Warning>

Select an available phone number from the dropdown. If none are available:

* Release a number from another agent
* Request new numbers via the "Request more numbers" link

### Language Model (LLM)

Select the AI model that powers your agent's conversations. Available models:

#### Anthropic (Claude)

| Model                | Features                                        |
| -------------------- | ----------------------------------------------- |
| **Claude Haiku 4.5** | Recommended. Ultra Fast. Best for tool calling. |

#### OpenAI (GPT)

| Model            | Features                   |
| ---------------- | -------------------------- |
| **GPT-4o**       | Balanced speed and quality |
| **GPT-4.1**      | Latest full model          |
| **GPT-4.1 Mini** | Faster, cost-effective     |
| **GPT-4.1 Nano** | Fastest, most economical   |

#### Google (Gemini)

| Model                      | Features                     |
| -------------------------- | ---------------------------- |
| **Gemini 3 Flash Preview** | Newest. Preview release.     |
| **Gemini 2.5 Flash**       | Fast and capable             |
| **Gemini 2.5 Flash Lite**  | Lighter, faster version      |
| **Gemini 2.0 Flash**       | Stable, fast                 |
| **Gemini 2.0 Flash Lite**  | Lightweight version          |
| **Gemini 1.5 Pro**         | Most capable, higher latency |
| **Gemini 1.5 Flash**       | Ultra Fast. Good balance.    |

<Info>
  For most use cases, start with **Claude Haiku 4.5** or **Gemini 2.5 Flash**. They offer the best balance of speed, quality, and cost.
</Info>

### Call Recording

Toggle to enable/disable call recordings. When enabled:

* All calls are recorded and stored
* Recordings are available in the Calls section
* Required for quality assurance and training

### Meetzy Foundation

Optional base components that enhance your agent:

| Component                  | Description                                 |
| -------------------------- | ------------------------------------------- |
| **Safety Guardrails**      | Prevents harmful or inappropriate responses |
| **Conversational Aspects** | Improves natural conversation flow          |
| **Agent Work Experience**  | Pre-trained behaviors for common scenarios  |

## Step 2: Greetings

Click the **Greetings** icon (Hand Raised) to configure welcome messages.

### Inbound Greeting

The first message when a customer calls your agent:

```
Hello! Thank you for calling {{company_name}}. How can I help you today?
```

### Outbound Greeting

The first message when your agent calls a customer:

```
Hi {{customer_name}}, this is {{agent_name}} from {{company_name}}. 
I'm calling regarding {{reason}}. Is this a good time to talk?
```

### Templates

Select from pre-built greeting templates:

* Sales Call
* Customer Support
* Appointment Reminder
* Survey Introduction

### Dynamic Variables

Use variables to personalize greetings:

| Variable               | Example             |
| ---------------------- | ------------------- |
| `{{customer_name}}`    | "John"              |
| `{{company_name}}`     | "Acme Corp"         |
| `{{agent_name}}`       | "Sarah"             |
| `{{appointment_date}}` | "March 15th at 2pm" |

## Step 3: System Prompt

The central editor contains the **System Prompt** - the core instructions that define your agent's behavior.

<Info>
  If you used prompt-first creation, Meetzy Copilot has already generated an optimized system prompt based on your description. You can review and customize it here.
</Info>

### Structure

A good system prompt includes:

1. **Role Definition**: Who the agent is
2. **Objective**: What the agent should accomplish
3. **Guidelines**: How to handle specific situations
4. **Constraints**: What the agent should NOT do

### Example

```markdown theme={null}
You are a friendly sales qualification agent for Acme Corp.

## Objective
Qualify leads by understanding their needs and budget, 
then schedule a demo with a sales representative.

## Guidelines
- Be conversational and friendly
- Ask about their current challenges
- Understand their timeline and budget
- If qualified, offer to schedule a demo

## Constraints
- Never discuss pricing directly
- Don't make promises about features
- Transfer to human if customer is upset
```

## Step 4: Input Params

Click the **Input Params** icon (Document) to define variables passed to each call.

Common parameters:

| Parameter     | Type   | Description              |
| ------------- | ------ | ------------------------ |
| `phone`       | String | Customer's phone number  |
| `name`        | String | Customer's name          |
| `email`       | String | Customer's email         |
| `company`     | String | Customer's company       |
| `lead_source` | String | Where the lead came from |

These parameters are available in your system prompt and greetings using `{{param_name}}` syntax.

## Step 5: Knowledge Base

Click the **Knowledge Base** icon (Book Stack) to add reference documents.

Upload documents the agent can reference during calls:

* Product FAQs
* Pricing information
* Company policies
* Common objections and responses

<Warning>
  The agent searches the knowledge base contextually during conversations. It doesn't memorize everything - it retrieves relevant information when needed.
</Warning>

## Step 6: Actions & Tools

Click the **Actions & Tools** icon (Wrench) to configure integrations.

Available actions:

| Action               | Description                   |
| -------------------- | ----------------------------- |
| **Transfer Call**    | Route to human agent          |
| **Send SMS**         | Send text message to customer |
| **Send Email**       | Send email to customer        |
| **Calendar Booking** | Schedule appointments         |
| **CRM Update**       | Update customer records       |
| **Custom Webhook**   | Call any external API         |

## Step 7: Evaluations

Click the **Evaluations** icon (Check Badge) to set up automated tests.

Create test scenarios to validate your agent:

```
Scenario: Customer asks about pricing
Expected: Agent should NOT discuss specific prices
Expected: Agent should offer to connect with sales team
```

## Step 8: Output Settings

Click the **Output Settings** icon (Inbox) to define data extraction.

Specify what information the agent should capture:

| Field          | Type    | Description              |
| -------------- | ------- | ------------------------ |
| `qualified`    | Boolean | Is the lead qualified?   |
| `budget_range` | String  | Estimated budget         |
| `timeline`     | String  | Purchase timeline        |
| `next_steps`   | String  | Agreed follow-up actions |
| `pain_points`  | Array   | Customer's challenges    |

Extracted data is available in call summaries and can be sent to your CRM.

## Publishing Your Agent

When your configuration is complete:

<Steps>
  <Step title="Validate">
    Click **Check** to validate your configuration. Fix any errors.
  </Step>

  <Step title="Test">
    Use the Playground (right panel) to test conversations.
  </Step>

  <Step title="Make test calls">
    Use "Init Call" from the agents list to make real phone calls.
  </Step>

  <Step title="Publish">
    Click **Publish** to deploy your agent to production.
  </Step>
</Steps>

<Warning>
  Publishing creates a new version. The previous version remains available for rollback if needed.
</Warning>

## Testing Your Agent

Before going live, thoroughly test your agent:

### Playground Testing

Use the right panel's Playground to:

* Simulate conversations
* Test different scenarios
* Verify response quality
* Check knowledge base integration

### Test Calls

Make actual phone calls to verify:

* Audio quality
* Response timing
* Phone number routing
* End-to-end functionality

## Next Steps

<CardGroup>
  <Card title="Manage Agents" icon="list" href="/agents/manage-agents">
    Learn to view, edit, and organize your agents.
  </Card>

  <Card title="Phone Health" icon="heart-pulse" href="/agents/phone-health">
    Monitor your phone numbers' health.
  </Card>
</CardGroup>


# Introduction to Meetzy
Source: https://docs.meetzy.es/getting-started/introduction

Discover Meetzy, the AI-powered voice agent platform for automating phone conversations

## What is Meetzy?

Meetzy is an AI-powered voice agent platform that enables businesses to automate phone conversations with natural, human-like interactions. Whether you need to handle inbound customer service calls or make outbound sales calls, Meetzy agents can manage conversations 24/7.

## Getting Started Options

Meetzy offers two ways to get started:

<CardGroup>
  <Card title="Self-Service Onboarding" icon="rocket" href="https://www.meetzy.io">
    Create your agent in minutes directly from our website with our new streamlined onboarding flow.
  </Card>

  <Card title="Traditional Login" icon="user" href="https://dashboard.meetzy.io">
    Access the full dashboard with advanced configuration and management features.
  </Card>
</CardGroup>

<Info>
  The self-service onboarding flow allows you to quickly create an agent by simply describing what you want it to do. The system will automatically configure your agent and guide you through the setup process.
</Info>

## Main Use Cases

<CardGroup>
  <Card title="Sales & Lead Qualification" icon="phone-arrow-up-right">
    Automatically call and qualify leads, schedule appointments, and follow up with prospects.
  </Card>

  <Card title="Customer Support" icon="headset">
    Handle inbound support calls, answer FAQs, and route complex issues to human agents.
  </Card>

  <Card title="Appointment Reminders" icon="calendar-check">
    Call customers to confirm appointments, send reminders, and handle rescheduling.
  </Card>

  <Card title="Surveys & Feedback" icon="clipboard-list">
    Conduct phone surveys, collect feedback, and gather customer insights.
  </Card>
</CardGroup>

## Platform Architecture

Meetzy consists of three main components:

| Component     | Description                                                                    |
| ------------- | ------------------------------------------------------------------------------ |
| **Dashboard** | Web interface to create, configure, and monitor your voice agents              |
| **AI Engine** | Processes conversations using LLMs (Claude, GPT, Gemini) for natural responses |
| **Telephony** | Handles phone calls through Twilio integration                                 |

## Main Dashboard Sections

| Section             | Description                       |
| ------------------- | --------------------------------- |
| 📊 **Insights**     | Analytics and performance metrics |
| 📣 **Marketing**    | Campaign management               |
| 📞 **Calls**        | Call history and recordings       |
| 🤖 **Agents**       | AI agent configuration            |
| 🔌 **Integrations** | Third-party connections           |
| 📖 **Docs**         | Documentation access              |

## Key Concepts

<AccordionGroup>
  <Accordion title="Agent" icon="robot">
    An AI-powered voice assistant that can make and receive phone calls. Each agent has its own configuration, personality, and capabilities.
  </Accordion>

  <Accordion title="Call" icon="phone">
    A phone conversation between an agent and a human. Calls can be inbound (customer calls you) or outbound (agent calls customer).
  </Accordion>

  <Accordion title="CRM" icon="users">
    Contact management system to store customer information and call history.
  </Accordion>

  <Accordion title="Knowledge Base" icon="book">
    Documents and FAQs that the agent can reference during conversations.
  </Accordion>

  <Accordion title="Actions & Tools" icon="wrench">
    External integrations and APIs the agent can call during conversations.
  </Accordion>
</AccordionGroup>

## Access Methods

<Tabs>
  <Tab title="Self-Service Onboarding">
    <Steps>
      <Step title="Visit meetzy.io">
        Go to our website and use the agent builder to describe your needs.
      </Step>

      <Step title="Quick authentication">
        Sign in with Google or Microsoft for secure access.
      </Step>

      <Step title="Auto-generated agent">
        Your agent is created automatically based on your description.
      </Step>

      <Step title="Test and deploy">
        Make test calls and publish when ready.
      </Step>
    </Steps>
  </Tab>

  <Tab title="Traditional Dashboard">
    <Steps>
      <Step title="Create your account">
        Sign up at [dashboard.meetzy.io](https://dashboard.meetzy.io) and complete your profile.
      </Step>

      <Step title="Create your first agent">
        Use the Advanced Editor to configure your AI voice agent.
      </Step>

      <Step title="Assign a phone number">
        Connect a phone number to your agent for inbound and outbound calls.
      </Step>

      <Step title="Test and publish">
        Make test calls to validate your agent, then publish to production.
      </Step>
    </Steps>
  </Tab>
</Tabs>

<Warning>
  If you start with the self-service onboarding, your agent configuration will be automatically transferred when you access the full dashboard for the first time.
</Warning>

## Authentication Options

Meetzy supports multiple authentication methods for secure access:

* **Google OAuth** - Sign in with your Google account
* **Microsoft OAuth** - Sign in with your Microsoft account
* **Email/Password** - Traditional login for existing accounts
* **Two-Factor Authentication** - Additional security for admin accounts

## Next Steps

<CardGroup>
  <Card title="Create Your First Agent" icon="plus" href="/getting-started/create-agent">
    Step-by-step guide to creating and configuring your first voice agent.
  </Card>

  <Card title="Manage Agents" icon="list" href="/agents/manage-agents">
    Learn how to view, edit, and organize your agents.
  </Card>
</CardGroup>


# Choose a Voice
Source: https://docs.meetzy.es/voices/choose-voice

Browse and select from the available voice catalog for your AI agents

## Voice Library Overview

The Voice Library is where you browse, preview, and select voices for your AI agents. Each voice has unique characteristics including gender, language support, and tonal qualities that affect how your agent sounds to callers.

<Info>
  The right voice significantly impacts caller perception and engagement. Take time to find one that matches your brand and use case.
</Info>

## Accessing the Voice Library

Navigate to **Voices** in the main navigation menu to access the Voice Library.

## Browsing Voices

The Voice Library displays all available voices in a table format:

| Column       | Description              |
| ------------ | ------------------------ |
| ID           | Unique voice identifier  |
| Name         | Voice name               |
| Gender       | Male, Female, or Neutral |
| Languages    | Supported languages      |
| Sample       | Play a sample            |
| Custom Audio | Generate custom samples  |

### Searching Voices

Use the search bar to filter voices by:

* Name
* Description
* Language
* Labels

<Tip>
  Search for specific characteristics like "warm", "professional", or "friendly" to find voices that match your needs.
</Tip>

## Previewing Voices

### Playing Samples

Each voice has a pre-recorded sample:

<Steps>
  <Step title="Find the Voice">
    Locate the voice you want to preview.
  </Step>

  <Step title="Click Play">
    Click the "▶️ Play" button in the Sample column.
  </Step>

  <Step title="Listen">
    The sample plays through your speakers.
  </Step>

  <Step title="Stop">
    Click "⏹️ Stop" to end playback.
  </Step>
</Steps>

### Generating Custom Audio

Test voices with your own text:

<Steps>
  <Step title="Enter Text">
    Type your custom text in the "Custom Text for Audio Generation" field.
  </Step>

  <Step title="Click Generate">
    Click "Generate" on the voice row you want to test.
  </Step>

  <Step title="Play">
    Once generated, click "▶️ Play" to hear the result.
  </Step>

  <Step title="Download (Optional)">
    Click "⬇️ Download" to save the audio file.
  </Step>
</Steps>

<Tip>
  Use a sample greeting or common phrase from your use case to get an accurate preview of how the voice will sound.
</Tip>

## Voice Settings

Fine-tune voice characteristics using the settings panel:

### Speed

| Range | Description   |
| ----- | ------------- |
| 0.70x | Slower speech |
| 1.00x | Normal speed  |
| 1.20x | Faster speech |

**Best practice:** Start with 1.00x and adjust based on testing.

### Naturalness (Stability)

| Value | Description                     |
| ----- | ------------------------------- |
| 0.0   | More natural, slight variations |
| 0.5   | Balanced                        |
| 1.0   | More consistent, less natural   |

**Best practice:** Lower values (0.2-0.4) typically sound more human.

### Consistency (Similarity Boost)

| Value | Description         |
| ----- | ------------------- |
| 0.0   | More variation      |
| 0.5   | Balanced            |
| 1.0   | Maximum consistency |

<Warning>
  High consistency values may cause audio artifacts. Test thoroughly before using values above 0.7.
</Warning>

## Selecting a Voice for Your Agent

Once you've found the right voice:

<Steps>
  <Step title="Note the Voice ID or Name">
    Remember or copy the voice identifier.
  </Step>

  <Step title="Open Advanced Editor">
    Navigate to your agent in the Advanced Editor.
  </Step>

  <Step title="Go to Basic Settings">
    Open the Basic Settings section.
  </Step>

  <Step title="Select Voice">
    Choose the voice from the dropdown.
  </Step>

  <Step title="Save">
    Save your agent configuration.
  </Step>
</Steps>

## Voice Characteristics

### By Gender

<Tabs>
  <Tab title="Female">
    * Generally perceived as friendly and approachable
    * Popular for customer service and support
    * Works well for healthcare and wellness
  </Tab>

  <Tab title="Male">
    * Often perceived as authoritative
    * Popular for finance and legal
    * Works well for technical support
  </Tab>

  <Tab title="Neutral">
    * Gender-neutral option
    * Works for diverse audiences
    * Modern and inclusive
  </Tab>
</Tabs>

### By Use Case

| Use Case         | Recommended Characteristics      |
| ---------------- | -------------------------------- |
| Customer Support | Warm, patient, clear             |
| Sales            | Energetic, persuasive, confident |
| Healthcare       | Calm, reassuring, professional   |
| Finance          | Trustworthy, professional, clear |
| Technology       | Modern, clear, knowledgeable     |

## Language Support

Voices may support one or multiple languages. Check the "Languages" column to ensure compatibility with your agent's language setting.

**Supported Languages include:**

* 🇺🇸 English (US)
* 🇬🇧 English (UK)
* 🇪🇸 Spanish
* 🇫🇷 French
* 🇩🇪 German
* 🇮🇹 Italian
* 🇵🇹 Portuguese
* And more...

<Warning>
  Using a voice with a language it doesn't support will result in poor quality or incorrect pronunciation.
</Warning>

## Best Practices

<AccordionGroup>
  <Accordion title="Match Brand Voice" icon="wand-magic-sparkles">
    Choose a voice that aligns with your brand personality. A luxury brand might want a sophisticated voice, while a tech startup might prefer a modern, casual one.
  </Accordion>

  <Accordion title="Consider Your Audience" icon="users">
    Think about who will be receiving calls. Different demographics may respond better to different voice characteristics.
  </Accordion>

  <Accordion title="Test with Real Content" icon="flask">
    Generate custom audio using actual phrases from your scripts to hear how the voice handles your specific content.
  </Accordion>

  <Accordion title="Check Language Quality" icon="language">
    If using non-English languages, test pronunciation of common terms and names.
  </Accordion>

  <Accordion title="Adjust Settings" icon="sliders">
    Fine-tune speed and naturalness to match your use case. Support calls might benefit from slower, clearer speech.
  </Accordion>
</AccordionGroup>

## Next Steps

<CardGroup>
  <Card title="Voice Designer" icon="palette" href="/voices/voice-designer">
    Create custom voices
  </Card>

  <Card title="Basic Settings" icon="gear" href="/advanced-editor/basic-settings">
    Configure agent voice settings
  </Card>
</CardGroup>


# Voice Designer
Source: https://docs.meetzy.es/voices/voice-designer

Create custom voices tailored to your specific needs

## What is Voice Designer?

Voice Designer is Meetzy's AI-powered tool for creating custom voices. Instead of choosing from the existing catalog, you can describe the exact voice characteristics you want, and the system generates multiple options for you to choose from.

<Info>
  Voice Designer uses advanced AI to create unique voices based on your text description.
</Info>

## When to Use Voice Designer

| Scenario                     | Recommendation     |
| ---------------------------- | ------------------ |
| Standard voice works fine    | Use Voice Library  |
| Need specific accent or tone | Use Voice Designer |
| Brand-specific voice         | Use Voice Designer |
| Custom character voice       | Use Voice Designer |
| Quick setup                  | Use Voice Library  |

## Creating a Custom Voice

### Step 1: Open Voice Designer

<Steps>
  <Step title="Go to Voice Library">
    Navigate to the Voices section.
  </Step>

  <Step title="Click Add New Voice">
    Click the "➕ Add New Voice" button.
  </Step>

  <Step title="Voice Designer Opens">
    The Voice Designer modal appears.
  </Step>
</Steps>

### Step 2: Describe Your Voice

Write a detailed description of the voice you want:

**Good Description Example:**

```
A warm, friendly female voice with a slight British accent, 
suitable for customer service. Should sound professional 
but approachable, with a calm and reassuring tone.
```

**Required Length:** 20-1000 characters

<Tip>
  Be specific about gender, accent, tone, and purpose for the best results.
</Tip>

### Description Best Practices

Include these elements:

| Element         | Examples                         |
| --------------- | -------------------------------- |
| Gender          | Female, Male, Neutral            |
| Accent          | British, American, Spanish       |
| Tone            | Warm, Professional, Energetic    |
| Characteristics | Calm, Confident, Friendly        |
| Use Case        | Customer service, Sales, Support |

### Enhance Description

Let AI improve your description:

<Steps>
  <Step title="Write Initial Description">
    Enter your basic voice description.
  </Step>

  <Step title="Click Enhance">
    Click the "⚡ Enhance Description" button.
  </Step>

  <Step title="Review">
    Review the enhanced description.
  </Step>

  <Step title="Edit if Needed">
    Make any final adjustments.
  </Step>
</Steps>

### Step 3: Add Sample Text

Provide text for generating voice samples:

**Good Sample Text:**

```
Hello, thank you for calling Acme Corporation. My name is 
Sarah and I'll be happy to assist you today. How may I 
help you with your account?
```

**Required Length:** 100-1000 characters

<Info>
  Use text that represents how your agent will actually speak. This helps you evaluate if the voice is right for your use case.
</Info>

### Generate Sample Text

Let AI create sample text based on your description:

<Steps>
  <Step title="Click Generate Sample Text">
    Click the "📝 Generate Sample Text" button.
  </Step>

  <Step title="Review">
    Review the generated text.
  </Step>

  <Step title="Customize">
    Modify as needed for your use case.
  </Step>
</Steps>

### Step 4: Generate Voices

<Steps>
  <Step title="Click Generate Voices">
    Click the "Generate Voices" button.
  </Step>

  <Step title="Wait">
    Voice generation takes 1-2 minutes.
  </Step>

  <Step title="Review Options">
    Multiple voice options are generated.
  </Step>
</Steps>

## Selecting a Voice

After generation, you'll see multiple voice options:

### Listen to Options

<Steps>
  <Step title="Play Each Option">
    Click the play button on each voice.
  </Step>

  <Step title="Compare">
    Listen to all options to compare.
  </Step>

  <Step title="Select Favorite">
    Click the radio button for your preferred voice.
  </Step>
</Steps>

### Waveform Visualization

When you select a voice, you'll see an audio waveform that animates during playback, helping you visualize the voice characteristics.

## Saving Your Voice

Once you've selected a voice:

<Steps>
  <Step title="Click Continue">
    Click "Continue" after selecting your voice.
  </Step>

  <Step title="Enter Voice Name">
    Give your voice a descriptive name.
  </Step>

  <Step title="Add Description">
    Write a brief description (optional but recommended).
  </Step>

  <Step title="Select Gender">
    Choose Male, Female, or Neutral.
  </Step>

  <Step title="Select Language">
    Choose the primary language.
  </Step>

  <Step title="Add Labels">
    Enter comma-separated labels for categorization.
  </Step>

  <Step title="Save">
    Click "Save Voice" to add to your library.
  </Step>
</Steps>

### Voice Properties

| Field       | Required | Description                          |
| ----------- | -------- | ------------------------------------ |
| Name        | Yes      | Unique voice name                    |
| Description | No       | Voice characteristics (min 20 chars) |
| Gender      | Yes      | Male, Female, or Neutral             |
| Language    | Yes      | Primary language                     |
| Labels      | No       | Comma-separated tags                 |

### Label Examples

Use labels to organize your custom voices:

```
professional, customer-service, warm
```

```
sales, energetic, american
```

```
support, calm, british
```

## Managing Custom Voices

### Finding Custom Voices

Custom voices appear in the Voice Library alongside pre-built voices. Use the search or filter by your labels to find them quickly.

### Using Custom Voices

1. Note the voice name or ID
2. Open your agent in Advanced Editor
3. Go to Basic Settings
4. Select your custom voice from the dropdown
5. Save

## Best Practices

<AccordionGroup>
  <Accordion title="Be Descriptive" icon="align-left">
    The more detailed your description, the better the AI can match your requirements. Include specific adjectives and characteristics.
  </Accordion>

  <Accordion title="Use Realistic Sample Text" icon="message">
    Provide sample text that represents actual phrases your agent will speak. This helps you evaluate voice quality in context.
  </Accordion>

  <Accordion title="Compare All Options" icon="ear-listen">
    Always listen to all generated options before selecting. Small differences can be significant in production.
  </Accordion>

  <Accordion title="Test in Context" icon="flask">
    After saving, test the voice in the Playground with various conversation scenarios.
  </Accordion>

  <Accordion title="Use Meaningful Labels" icon="tags">
    Good labels make it easier to find and manage voices across multiple agents and use cases.
  </Accordion>
</AccordionGroup>

## Troubleshooting

| Issue                          | Solution                                   |
| ------------------------------ | ------------------------------------------ |
| Generation takes too long      | Wait up to 2 minutes; can cancel and retry |
| Voices don't match description | Try a more detailed description            |
| Poor quality audio             | Adjust sample text length and content      |
| Can't find saved voice         | Check labels and use search                |

## Example Voice Descriptions

### Professional Support

```
A professional female voice with an American accent. 
Calm, patient, and reassuring tone suitable for 
customer support. Should sound knowledgeable but 
approachable.
```

### Energetic Sales

```
An energetic male voice with enthusiasm and confidence. 
Slightly faster pace, engaging and persuasive. 
American accent, suitable for sales calls.
```

### Sophisticated Concierge

```
An elegant female voice with a subtle British accent. 
Refined and polished, suitable for luxury hospitality 
or high-end services. Warm but sophisticated.
```

## Next Steps

<CardGroup>
  <Card title="Voice Library" icon="microphone" href="/voices/choose-voice">
    Browse all available voices
  </Card>

  <Card title="Basic Settings" icon="gear" href="/advanced-editor/basic-settings">
    Apply voice to your agent
  </Card>
</CardGroup>