### startScreenSharing Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Example of starting screen sharing with display media options.

```typescript
vapi.startScreenSharing({
  video: {
    mandatory: {
      chromeMediaSource: 'screen'
    }
  }
});
```

--------------------------------

### Start Method Examples

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Illustrates various ways to use the start() method, including with assistant IDs, assistant objects, overrides, and custom room settings.

```typescript
// Start with assistant ID
const call = await vapi.start('assistant-id-123');

// Start with assistant object
const call = await vapi.start({
  model: {
    provider: 'openai',
    model: 'gpt-3.5-turbo',
    messages: [
      {
        role: 'system',
        content: 'You are a helpful assistant.'
      }
    ]
  },
  voice: {
    provider: '11labs',
    voiceId: 'burt'
  }
});

// Start with overrides to set variables
const call = await vapi.start(
  'assistant-id-123',
  {
    variableValues: { name: 'John', age: '30' },
    recordingEnabled: false
  }
);

// Start with custom room settings
const call = await vapi.start(
  'assistant-id-123',
  undefined,
  undefined,
  undefined,
  undefined,
  { roomDeleteOnUserLeaveEnabled: false }
);
```

--------------------------------

### startAudioOff Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Whether to start with audio/microphone disabled.

```typescript
const vapi = new Vapi('api-token', undefined, undefined, {
  startAudioOff: true // Start muted
});
```

--------------------------------

### Example Usage of start() and reconnect()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Demonstrates how to start a new call and then reconnect to an existing call using the Vapi SDK.

```typescript
// Store the call object during initial start
let currentCall: Call | null = null;

currentCall = await vapi.start('assistant-id', undefined, undefined, undefined, undefined, {
  roomDeleteOnUserLeaveEnabled: false
});

// Later, reconnect to the same call
if (currentCall) {
  await vapi.reconnect(currentCall as any);
}
```

--------------------------------

### Complete Type Usage Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

A comprehensive example showcasing the creation of an assistant, starting a call with overrides, and sending a typed message using the Vapi client SDK.

```typescript
import Vapi, {
  client,
  CreateAssistantDTO,
  Assistant,
  CreateCallDTO,
  Call,
  AssistantOverrides,
  VapiClientToServerMessage
} from '@vapi-ai/web';

// Create assistant type
const assistantConfig: CreateAssistantDTO = {
  name: 'Support Bot',
  model: { provider: 'openai', model: 'gpt-4', messages: [] },
  voice: { provider: '11labs', voiceId: 'adam' }
};

// Create assistant
const assistant: Assistant = await client.assistant.assistantControllerCreate(
  assistantConfig
);

// Start call with overrides
const overrides: AssistantOverrides = {
  variableValues: { name: 'John' }
};

const vapi = new Vapi('api-key');
const call = await vapi.start(assistant.id, overrides);

// Send typed message
const msg: VapiClientToServerMessage = {
  type: 'say',
  message: 'Hello!'
};

vapi.send(msg);
```

--------------------------------

### Start a new call with assistant configuration

Source: https://github.com/vapiai/client-sdk-web/blob/main/README.md

You can start a new call by calling the `start` method and passing an `assistant` object or `assistantId`.

```javascript
vapi.start({
  model: {
    provider: "openai",
    model: "gpt-3.5-turbo",
    messages: [
      {
        role: "system",
        content: "You are an assistant.",
      },
     ],
   },
   voice: {
    provider: "11labs",
    voiceId: "burt",
  },
  ...
});
```

--------------------------------

### Example GET /v2/call/export

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/endpoints.md

Example of how to export calls with specified format, async option, and limit.

```http
GET /v2/call/export?format=csv&async=false&limit=1000 HTTP/1.1
Authorization: Bearer api_token
```

--------------------------------

### Installation

Source: https://github.com/vapiai/client-sdk-web/blob/main/README.md

You can install the package via npm.

```bash
npm install @vapi-ai/web
```

--------------------------------

### Pattern 1: Full Setup

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Import both the Vapi class and the client instance for a full setup.

```typescript
import Vapi, { client } from '@vapi-ai/web';

const vapi = new Vapi('api-key');
const assistants = await client.assistant.assistantControllerFindAll();
```

--------------------------------

### Start a new call with assistant ID

Source: https://github.com/vapiai/client-sdk-web/blob/main/README.md

You can start a new call by calling the `start` method and passing an `assistant` object or `assistantId`.

```javascript
vapi.start('your-assistant-id');
```

--------------------------------

### List Assistants Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/endpoints.md

Example query parameters for listing assistants.

```http
GET /assistant?limit=50&createdAtGe=2024-01-01T00:00:00Z HTTP/1.1
Authorization: Bearer api_token
```

--------------------------------

### Create Assistant Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/endpoints.md

Example request body for creating a new assistant.

```typescript
POST /assistant HTTP/1.1
Authorization: Bearer api_token
Content-Type: application/json

{
  "name": "Customer Support Bot",
  "firstMessage": "Hello, how can I help?",
  "model": {
    "provider": "openai",
    "model": "gpt-3.5-turbo",
    "messages": [
      {
        "role": "system",
        "content": "You are a helpful assistant."
      }
    ]
  },
  "voice": {
    "provider": "11labs",
    "voiceId": "burt"
  }
}
```

--------------------------------

### Recommended Practices: Do This

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/daily-guards.md

Examples of recommended practices for configuring audio sources and permission prompts with Vapi.

```typescript
// Use default audio
const vapi = new Vapi('api-token', undefined, undefined, {
  audioSource: true
});

// Use specific device
await vapi.setInputDevicesAsync({
  audioSource: { deviceId: deviceId }
});

// Use microphone-required permission prompt
const vapi = new Vapi('api-token', undefined, {
  alwaysIncludeMicInPermissionPrompt: true
});
```

--------------------------------

### Authentication Setup

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Shows how to set the API token for authentication using the client SDK.

```typescript
import { client } from '@vapi-ai/web';

// Set authentication
client.setSecurityData('your-api-token');

// Now all requests include Authorization: Bearer <token>
```

--------------------------------

### runNetworkTestsStandalone Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Example of performing network connectivity tests.

```typescript
const results = await Vapi.runNetworkTestsStandalone();
console.log('Network tests:', results);
```

--------------------------------

### Event Examples

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Examples of how to listen to various Vapi events, including call lifecycle, progress tracking, messages, speech activity, errors, and network monitoring.

```typescript
// Listen to call lifecycle
vapi.on('call-start', () => {
  console.log('Call started!');
});

vapi.on('call-end', () => {
  console.log('Call ended');
});

// Track progress during initialization
vapi.on('call-start-progress', (event) => {
  console.log(`${event.stage}: ${event.status}`, event.metadata);
});

// Listen for assistant messages
vapi.on('message', (message) => {
  console.log('Assistant said:', message);
});

// Track speech activity
vapi.on('speech-start', () => {
  console.log('User is speaking');
});

vapi.on('speech-end', () => {
  console.log('User stopped speaking');
});

vapi.on('volume-level', (volume) => {
  console.log(`Volume: ${Math.round(volume * 100)}%`);
});

// Error handling
vapi.on('error', (error) => {
  console.error('Error:', error);
});

vapi.on('camera-error', (error) => {
  console.error('Camera error:', error);
});

// Network monitoring
vapi.on('network-quality-change', (event) => {
  console.log('Network quality:', event);
});
```

--------------------------------

### increaseMicLevel Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Example of increasing the microphone level.

```typescript
await vapi.increaseMicLevel(1.5); // Increase volume by 1.5x
```

--------------------------------

### Constructor Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Demonstrates how to instantiate the Vapi class with and without custom configuration.

```typescript
import Vapi from '@vapi-ai/web';

const vapi = new Vapi('your-public-api-key');

// With custom configuration
const vapiCustom = new Vapi(
  'your-public-api-key',
  'https://api.vapi.ai',
  {
    alwaysIncludeMicInPermissionPrompt: true
  },
  {
    audioSource: true,
    startAudioOff: false
  }
);
```

--------------------------------

### Example Usage of setOutputDeviceAsync()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Shows how to set the audio output device using setOutputDeviceAsync().

```typescript
await vapi.setOutputDeviceAsync({
  deviceId: 'speaker-device-id'
});
```

--------------------------------

### Assistant Model Configuration Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Configures the language model for an assistant.

```typescript
{
  provider: 'openai',
  model: 'gpt-4-turbo',
  messages: [
    {
      role: 'system',
      content: 'You are a helpful support agent.'
    }
  ],
  temperature: 0.7,
  maxTokens: 1000
}
```

--------------------------------

### Testing Changes Locally

Source: https://github.com/vapiai/client-sdk-web/blob/main/RELEASE.md

Commands to build, test, and run the example project locally, as well as clean build artifacts.

```bash
npm run test:example

npm run pack:local

npm run dev:example

npm run clean-builds
```

--------------------------------

### Create Web Call Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/endpoints.md

Example of creating a web call using a POST request.

```typescript
POST /call/web HTTP/1.1
Authorization: Bearer api_token
Content-Type: application/json

{
  "assistantId": "assistant-123"
}
```

--------------------------------

### Stop Method Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

A simple example demonstrating how to stop the current call.

```typescript
await vapi.stop();
```

--------------------------------

### audioSource Example - Use default microphone

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Specifies the audio input device. `true`: Use default microphone

```typescript
const vapi = new Vapi('api-token', undefined, undefined, {
  audioSource: true
});
```

--------------------------------

### audioSource Example - Use specific device

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Specifies the audio input device. `string`: Device ID from `enumerateDevices()`

```typescript
// Use specific device
const devices = await navigator.mediaDevices.enumerateDevices();
const micDevice = devices.find(d => d.kind === 'audioinput');
const vapi = new Vapi('api-token', undefined, undefined, {
  audioSource: { deviceId: micDevice.deviceId }
});
```

--------------------------------

### Create Call Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/endpoints.md

Example request body for creating a new call.

```typescript
POST /call HTTP/1.1
Authorization: Bearer api_token
Content-Type: application/json

{
  "assistantId": "assistant-123",
  "customerNumber": "+1234567890",
  "phoneNumberId": "phone-456"
}
```

--------------------------------

### Example Usage of setInputDevicesAsync()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Demonstrates how to change the audio input device using setInputDevicesAsync().

```typescript
// Change to a specific audio device
const devices = await navigator.mediaDevices.enumerateDevices();
const audioDevices = devices.filter(d => d.kind === 'audioinput');
if (audioDevices.length > 0) {
  await vapi.setInputDevicesAsync({
    audioSource: { deviceId: audioDevices[0].deviceId }
  });
}
```

--------------------------------

### safeSetLocalAudio Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/daily-guards.md

Demonstrates the usage of safeSetLocalAudio for enabling and disabling the microphone.

```typescript
import { safeSetLocalAudio } from '@vapi-ai/web';

const dailyCall = vapi.getDailyCallObject();

// Enable microphone
safeSetLocalAudio(dailyCall, true);

// Disable microphone (mute)
safeSetLocalAudio(dailyCall, false);

// Handle null safely
try {
  safeSetLocalAudio(null, true);
} catch (error) {
  console.error('No active call');
}
```

--------------------------------

### Example Usage of safeSetInputDevicesAsync

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/daily-guards.md

Demonstrates how to use `safeSetInputDevicesAsync` to manage audio input devices, including switching to a specific device and handling unsafe `audioSource: false` configurations.

```typescript
import { safeSetInputDevicesAsync } from '@vapi-ai/web';

const dailyCall = vapi.getDailyCallObject();

// Get available audio devices
const devices = await navigator.mediaDevices.enumerateDevices();
const audioDevices = devices.filter(d => d.kind === 'audioinput');

// Switch to specific device
await safeSetInputDevicesAsync(dailyCall, {
  audioSource: { deviceId: audioDevices[0].deviceId }
});

// Start with audio enabled (safe)
await safeSetInputDevicesAsync(dailyCall, {
  audioSource: true
});

// Try to disable audio (gets converted to safe default)
await safeSetInputDevicesAsync(dailyCall, {
  audioSource: false  // Warning logged, property removed
});
// Console: "[Vapi] setInputDevicesAsync with audioSource:false detected. 
//           This can cause Chrome 140+ issues. Using default device instead."
```

--------------------------------

### audioSource Example - Use custom audio track

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Specifies the audio input device. `MediaStreamTrack`: Custom audio track

```typescript
// Use custom audio track
const stream = await navigator.mediaDevices.getUserMedia({ audio: true });
const track = stream.getAudioTracks()[0];
const vapi = new Vapi('api-token', undefined, undefined, {
  audioSource: track
});
```

--------------------------------

### Client Setup

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Initializes the Vapi API client with your API token and optionally sets the base URL.

```typescript
import { client } from '@vapi-ai/web';

// Set the base URL (defaults to https://api.vapi.ai)
client.baseUrl = 'https://api.vapi.ai';

// Set your API token for authentication
client.setSecurityData('your-api-token');
```

--------------------------------

### createSafeDailyConfig Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/daily-guards.md

Demonstrates the usage of createSafeDailyConfig with different inputs.

```typescript
import { createSafeDailyConfig } from '@vapi-ai/web';

// Safe configuration
const config1 = createSafeDailyConfig({
  alwaysIncludeMicInPermissionPrompt: true,
  avoidEval: true
});
// Returns: { alwaysIncludeMicInPermissionPrompt: true, avoidEval: true }

// Unsafe configuration (false value) — auto-corrected
const config2 = createSafeDailyConfig({
  alwaysIncludeMicInPermissionPrompt: false
});
// Logs warning and returns: { }
// Console: "[Vapi] alwaysIncludeMicInPermissionPrompt:false detected. 
//           This can cause Chrome 140+ issues. Removing the property."

// No config
const config3 = createSafeDailyConfig();
// Returns: { }
```

--------------------------------

### Variable Substitution Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Demonstrates how to use `variableValues` to replace placeholders in assistant messages during a call.

```typescript
const call = await vapi.start(
  'assistant-id',
  {
    variableValues: {
      'customer_name': 'John Doe',
      'order_id': 'ORD-12345',
      'account_balance': '$500.00'
    }
  }
);
```

--------------------------------

### on Event Method

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Example of subscribing to a Vapi event.

```typescript
on<E extends VapiEventNames>(event: E, listener: VapiEventListeners[E]): this
```

--------------------------------

### stopScreenSharing Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Example of stopping the active screen share.

```typescript
vapi.stopScreenSharing();
```

--------------------------------

### getDailyCallObject Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Example of retrieving the Daily.co call object.

```typescript
const dailyCall = vapi.getDailyCallObject();
if (dailyCall) {
  // Use Daily.co API directly
  const participants = dailyCall.participants();
}
```

--------------------------------

### Example Usage of say()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Demonstrates how to use the say() method to speak a message and optionally end the call.

```typescript
vapi.say("Thank you for calling. Goodbye!", true);
```

--------------------------------

### Basic Usage

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/README.md

Initialize the Vapi SDK, listen to events, start a call, control it, and stop it.

```typescript
import Vapi from '@vapi-ai/web';

// Initialize
const vapi = new Vapi('your-public-api-key');

// Listen to events
vapi.on('call-start', () => console.log('Call started'));
vapi.on('speech-start', () => console.log('User speaking'));
vapi.on('message', (msg) => console.log('Message:', msg));
vapi.on('error', (err) => console.error('Error:', err));

// Start a call
const call = await vapi.start('assistant-id-or-object');

// Control the call
vapi.setMuted(true);          // Mute user
vapi.say('Hello there!');     // Speak to user
vapi.send({                    // Send control message
  type: 'add-message',
  message: { role: 'system', content: 'User clicked button' }
});

// Stop the call
await vapi.stop();
```

--------------------------------

### Error Handling Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Demonstrates how to handle API errors using a try-catch block.

```typescript
try {
  const assistant = await client.assistant.assistantControllerCreate({
    model: { provider: 'openai', model: 'gpt-3.5-turbo' },
    voice: { provider: '11labs', voiceId: 'burt' }
  });
} catch (error: any) {
  console.error('API Error:', error.response?.status, error.message);
}
```

--------------------------------

### Retry Logic for Starting a Call

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

This pattern demonstrates how to implement retry logic with exponential backoff when starting a call, to handle transient errors.

```typescript
async function startWithRetry(
  assistantId: string,
  maxAttempts = 3
) {
  let lastError;
  
  for (let i = 0; i < maxAttempts; i++) {
    try {
      return await vapi.start(assistantId);
    } catch (error) {
      lastError = error;
      // Exponential backoff
      await new Promise(resolve => 
        setTimeout(resolve, Math.pow(2, i) * 1000)
      );
    }
  }
  
  throw lastError;
}

// Usage
try {
  const call = await startWithRetry('assistant-id');
} catch (error) {
  console.error('Failed after retries:', error);
}
```

--------------------------------

### LLM Error Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

An example structure for an LLM error, indicating a failure with the language model service.

```typescript
{
  endedReason: 'pipeline-error-openai-401-unauthorized'
  // or other LLM errors
}
```

--------------------------------

### Example Usage of createSafeDailyFactoryOptions

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/daily-guards.md

Illustrates how `createSafeDailyFactoryOptions` handles safe and unsafe audio source configurations, including auto-correction for `audioSource: false`.

```typescript
import { createSafeDailyFactoryOptions } from '@vapi-ai/web';

// Safe options (true or device spec)
const opts1 = createSafeDailyFactoryOptions({
  audioSource: true,
  startAudioOff: false
});
// Returns: { audioSource: true, startAudioOff: false }

// Unsafe options (false value) — auto-corrected
const opts2 = createSafeDailyFactoryOptions({
  audioSource: false,
  startAudioOff: true
});
// Logs warning and returns: { audioSource: true, startAudioOff: true }
// Console: "[Vapi] audioSource:false detected in factory options. 
//           This can cause Chrome 140+ issues. Defaulting to true."

// Device ID (safe)
const opts3 = createSafeDailyFactoryOptions({
  audioSource: { deviceId: 'device-123' }
});
// Returns: { audioSource: { deviceId: 'device-123' } }

// No options
const opts4 = createSafeDailyFactoryOptions();
// Returns: { }
```

--------------------------------

### StartCallOptions Type

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/types.md

Options for starting a call.

```typescript
type StartCallOptions = {
  roomDeleteOnUserLeaveEnabled?: boolean;
}
```

--------------------------------

### Example Usage of send()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Illustrates how to use the send() method with different message types: adding a system message, controlling the assistant, using text-to-speech, and ending the call.

```typescript
// Add a system message
vapi.send({
  type: 'add-message',
  message: {
    role: 'system',
    content: 'The user just clicked a button.'
  },
  triggerResponseEnabled: true
});

// Control the assistant
vapi.send({
  type: 'control',
  control: 'mute-assistant'
});

// Use text-to-speech
vapi.send({
  type: 'say',
  message: 'Hello, how can I help you today?',
  endCallAfterSpoken: false,
  interruptionsEnabled: true
});

// End the call
vapi.send({
  type: 'end-call'
});
```

--------------------------------

### Recommended Practices: Don't Do This

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/daily-guards.md

Examples of unsafe configurations to avoid when using Vapi, highlighting automatic corrections applied by the SDK.

```typescript
// ❌ Disable audio via false (unsafe)
const vapi = new Vapi('api-token', undefined, undefined, {
  audioSource: false  // Converted to true automatically
});

// ❌ Set permission prompt to false (unsafe)
const vapi = new Vapi('api-token', undefined, {
  alwaysIncludeMicInPermissionPrompt: false  // Removed automatically
});

// ❌ Direct Daily.co call with unsafe values
dailyCall.setInputDevicesAsync({
  audioSource: false  // Can cause issues - use safeSetInputDevicesAsync
});
```

--------------------------------

### Import Organization Examples

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Illustrates different ways to import from the Vapi client SDK, including default imports, named exports, and type-only imports.

```typescript
// Import default
import Vapi from '@vapi-ai/web';

// Import named exports
import { 
  client,
  // Types
  Assistant,
  Call,
  // Utilities
  createSafeDailyConfig
} from '@vapi-ai/web';

// Import types only (not at runtime)
import type { 
  CreateAssistantDTO,
  AssistantOverrides
} from '@vapi-ai/web';
```

--------------------------------

### Recording Override Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Shows how to disable call recording for a specific call using `recordingEnabled` in the overrides.

```typescript
const call = await vapi.start(
  'assistant-id',
  {
    recordingEnabled: false // Disable recording for this call
  }
);
```

--------------------------------

### avoidEval Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Disables `eval()` in Daily.co runtime. Recommended for security.

```typescript
const vapi = new Vapi(
  'api-token',
  undefined,
  { avoidEval: true }
);
```

--------------------------------

### getCampaign()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Gets a specific campaign by ID.

```typescript
client.campaign.campaignControllerFindOne(id: string): Promise<Campaign>
```

--------------------------------

### Default Configuration Summary

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Provides a summary of the default configuration settings for the Vapi client SDK, including API base URL, Daily.co call configuration, and start call options.

```typescript
{
  apiBaseUrl: 'https://api.vapi.ai',
  dailyCallConfig: {},
  dailyCallObject: {
    audioSource: true,
    startAudioOff: false
  },
  startCallOptions: {
    roomDeleteOnUserLeaveEnabled: true
  }
}
```

--------------------------------

### Error Recovery

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/README.md

Handles call start failures by logging the error, waiting for 2 seconds, and then retrying to start the call.

```typescript
vapi.on('call-start-failed', async (event) => {
  console.error(`Failed at ${event.stage}: ${event.error}`);
  // Retry logic
  await new Promise(r => setTimeout(r, 2000));
  await vapi.start('assistant-id');
});
```

--------------------------------

### Progress Tracking

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

Logs call start progress events and handles failures by logging and reporting them. It also analyzes failed stages when a call start fails.

```typescript
const progressLog: CallStartProgressEvent[] = [];

vapi.on('call-start-progress', (event) => {
  progressLog.push(event);
  
  if (event.status === 'failed') {
    console.error(`Failed at stage: ${event.stage}`, event);
    // Report to analytics
    reportFailure(event);
  }
});

vapi.on('call-start-failed', (event) => {
  // Analyze where it failed
  const failedStages = progressLog.filter(e => e.status === 'failed');
  console.error('Failed stages:', failedStages);
});
```

--------------------------------

### Call Start Progress Event

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

Logs the progress of call initialization stages.

```typescript
vapi.on('call-start-progress', (event: CallStartProgressEvent) => {
  console.log(`Stage: ${event.stage}, Status: ${event.status}`);
});
```

--------------------------------

### Example Usage of end()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Shows how to call the end() method to terminate the current call.

```typescript
vapi.end();
```

--------------------------------

### Index Signatures Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Illustrates the use of index signatures for flexible properties in types like SerializedError and AssistantOverrides.

```typescript
interface SerializedError {
  message: string;
  name?: string;
  [key: string]: any;  // Allows additional properties
}

interface AssistantOverrides {
  variableValues?: Record<string, string>;  // Map of any variables
}
```

--------------------------------

### once Event Method

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Example of subscribing to a Vapi event that fires only once.

```typescript
once<E extends VapiEventNames>(event: E, listener: VapiEventListeners[E]): this
```

--------------------------------

### Nullable/Optional Properties Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Shows how response types often include optional properties to reflect potential nullability from the database, such as 'id', 'status', 'duration', and 'recordingUrl'.

```typescript
interface Call {
  id?: string;           // Optional ID from server
  status?: string;       // Optional status
  duration?: number;     // Optional duration
  recordingUrl?: string; // May not have recording
  // ... many optional properties
}
```

--------------------------------

### Call Start Failed Event

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

Logs an error message when call initialization fails.

```typescript
vapi.on('call-start-failed', (event: CallStartFailedEvent) => {
  console.error(`Failed at stage: ${event.stage}, Error: ${event.error}`);
});
```

--------------------------------

### alwaysIncludeMicInPermissionPrompt Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Ensures microphone permission is always requested, even if audio is disabled initially.

```typescript
const vapi = new Vapi(
  'api-token',
  undefined,
  { alwaysIncludeMicInPermissionPrompt: true }
);
```

--------------------------------

### Call Start Success Event

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

Logs a message when the call initialization completes successfully.

```typescript
vapi.on('call-start-success', (event: CallStartSuccessEvent) => {
  console.log('Call started successfully in', event.totalDuration, 'ms');
});
```

--------------------------------

### Conditional Types Example Usage

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Shows how type inference works with the 'on' method for different event types.

```typescript
vapi.on('speech-start', () => {
  // Correct: listener is () => void
});

vapi.on('volume-level', (volume: number) => {
  // Correct: listener is (volume: number) => void
});

vapi.on('message', (msg: any) => {
  // Correct: listener is (msg: any) => void
});
```

--------------------------------

### Get Assistant

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Retrieves a specific assistant by its ID.

```typescript
client.assistant.assistantControllerFindOne(id: string): Promise<Assistant>
```

--------------------------------

### Accessing the Daily.co Call Object

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Demonstrates how to retrieve the Daily.co call object to directly interact with its API for advanced customization, such as getting participant information or updating input settings.

```typescript
const dailyCall = vapi.getDailyCallObject();

if (dailyCall) {
  // Use Daily.co API directly
  const participants = dailyCall.participants();
  dailycall.updateInputSettings({ audio: { enabled: true } });
}
```

--------------------------------

### Example Usage of setMuted()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Demonstrates how to mute and unmute the user's microphone using the setMuted() method.

```typescript
vapi.setMuted(true); // Mute the microphone
vapi.setMuted(false); // Unmute the microphone
```

--------------------------------

### Vapi Constructor with API Key and URL

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Demonstrates how to initialize the Vapi SDK with an API key and API URL, falling back to environment variables or defaults.

```typescript
const vapi = new Vapi(
  process.env.VAPI_API_KEY || 'default-key',
  process.env.VAPI_API_URL || 'https://api.vapi.ai'
);
```

--------------------------------

### Example Usage of isMuted()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Shows how to retrieve the microphone's mute state using the isMuted() method.

```typescript
const muted = vapi.isMuted();
console.log('Microphone is muted:', muted);
```

--------------------------------

### Start call with assistant overrides

Source: https://github.com/vapiai/client-sdk-web/blob/main/README.md

You can override existing assistant parameters or set variables with the `assistant_overrides` parameter. Assume the first message is `Hey, {{name}} how are you?` and you want to set the value of `name` to `John`:

```javascript
const assistantOverrides = {
  recordingEnabled: false,
  variableValues: {
    name: 'John',
  },
};

vapi.start(
  'your-assistant-id',
  assistantOverrides,
);
```

--------------------------------

### Live Control

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/README.md

Listens for speech start and end events to manage a speaking indicator and demonstrates how to send control messages like 'mute-assistant' during a call.

```typescript
vapi.on('speech-start', () => {
  // User is speaking
  showSpeakingIndicator();
});

vapi.on('speech-end', () => {
  // User stopped speaking
  hideSpeakingIndicator();
});

// Control assistant during call
if (needsIntervention) {
  vapi.send({
    type: 'control',
    control: 'mute-assistant' // Silence the assistant
  });
}
```

--------------------------------

### Discriminated Unions Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Shows how discriminated unions provide type-safe handling of messages, with TypeScript inferring the correct structure based on the 'type' property.

```typescript
import type { VapiClientToServerMessage } from '@vapi-ai/web';

const msg: VapiClientToServerMessage = {
  type: 'say',  // TypeScript infers correct structure
  message: 'Hello',
  endCallAfterSpoken: true
};

// Type narrowing works
if (msg.type === 'add-message') {
  // msg.message exists here
  const m = msg.message;
}
```

--------------------------------

### Network Tests for Call Startup

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Shows how to run standalone network connectivity tests before initiating a Vapi call to ensure a stable connection.

```typescript
const results = await Vapi.runNetworkTestsStandalone();
if (results.networkConnectivity.result === 'success') {
  // Safe to start call
  await vapi.start('assistant-id');
} else {
  console.warn('Network connectivity issue:', results);
}
```

--------------------------------

### Transcription Error Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

An example structure for a transcription error, indicating a failure in the speech-to-text process.

```typescript
{
  endedReason: 'pipeline-error-deepgram-transcriber-failed'
  // or other transcriber errors
}
```

--------------------------------

### Main Entry Point

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Import and initialize the main Vapi class.

```typescript
import Vapi from '@vapi-ai/web';

// Initialize
const vapi = new Vapi('api-key');
```

--------------------------------

### Voice/TTS Error Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

An example structure for a voice or text-to-speech error, indicating a failure with the TTS provider.

```typescript
{
  endedReason: 'pipeline-error-eleven-labs-voice-failed'
  // or other voice provider errors
}
```

--------------------------------

### getChat()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Gets a specific chat by ID.

```typescript
client.chat.chatControllerGetChat(id: string): Promise<Chat>
```

--------------------------------

### getSquad()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Gets a specific squad by ID.

```typescript
client.squad.squadControllerFindOne(id: string): Promise<Squad>
```

--------------------------------

### Create Vapi instance

Source: https://github.com/vapiai/client-sdk-web/blob/main/README.md

Then, create a new instance of the Vapi class, passing your Public Key as a parameter to the constructor.

```javascript
const vapi = new Vapi('your-public-key');
```

--------------------------------

### Get Call

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Retrieves a specific call by its ID.

```typescript
client.call.callControllerFindOne(id: string): Promise<Call>
```

--------------------------------

### removeListener Event Method

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Example of removing an event listener.

```typescript
removeListener<E extends VapiEventNames>(event: E, listener: VapiEventListeners[E]): this
```

--------------------------------

### createCampaign()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Creates a new campaign.

```typescript
client.campaign.campaignControllerCreate(data: CreateCampaignDTO): Promise<Campaign>
```

--------------------------------

### removeAllListeners Event Method

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

Example of removing all listeners for a specific event or all events.

```typescript
removeAllListeners(event?: VapiEventNames): this
```

--------------------------------

### createWebChat()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Creates a web-based chat.

```typescript
client.chat.chatControllerCreateWebChat(data: CreateWebChatDTO): Promise<WebChat | CreateChatStreamResponse>
```

--------------------------------

### First Message Configuration

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Configures the initial message the assistant speaks or waits for, with options for immediate speech or waiting for user input.

```typescript
firstMessage?: string;
firstMessageMode?: 'speak-first-message' | 'wait-for-speech';
```

```typescript
{
  firstMessage: 'Hello! How can I help you today?',
  firstMessageMode: 'speak-first-message'
}
```

--------------------------------

### Constructor Signature

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

The Vapi class accepts configuration parameters to customize call behavior and audio handling.

```typescript
constructor(
  apiToken: string,
  apiBaseUrl?: string,
  dailyCallConfig?: Pick<DailyAdvancedConfig, 'avoidEval' | 'alwaysIncludeMicInPermissionPrompt'>,
  dailyCallObject?: Pick<DailyFactoryOptions, 'audioSource' | 'startAudioOff'>
)
```

--------------------------------

### Configuration Types

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Illustrates different configuration types available for Daily.co and Vapi factory options, as well as call options and assistant overrides.

```typescript
DailyAdvancedConfig
├── avoidEval
├── alwaysIncludeMicInPermissionPrompt
└── ... (other Daily.co options)

DailyFactoryOptions
├── audioSource
├── startAudioOff
└── ... (other factory options)

StartCallOptions
└── roomDeleteOnUserLeaveEnabled

AssistantOverrides
├── recordingEnabled
├── variableValues
├── modelOverrides
└── voiceOverrides
```

--------------------------------

### listCampaigns()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Lists all campaigns.

```typescript
client.campaign.campaignControllerFindAll(query?: {
  id?: string;
  status?: 'scheduled' | 'in-progress' | 'ended';
  page?: number;
  limit?: number;
  // ... more filters
}): Promise<CampaignPaginatedResponse>
```

--------------------------------

### Import Vapi class

Source: https://github.com/vapiai/client-sdk-web/blob/main/README.md

First, import the Vapi class from the package.

```javascript
import Vapi from '@vapi-ai/web';
```

--------------------------------

### createOpenAiWebChat()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Creates a web chat using OpenAI-compatible API.

```typescript
client.chat.chatControllerCreateOpenAiWebChat(data: OpenAIWebChatRequest): Promise<ResponseObject | ResponseTextDeltaEvent | ResponseTextDoneEvent | ResponseCompletedEvent | ResponseErrorEvent>
```

--------------------------------

### List Assistant Versions

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Lists all versions of a specific assistant, with options for pagination.

```typescript
client.assistant.assistantControllerFindVersions(id: string, query?: {
  page?: number;
  limit?: number;
  pageState?: string;
}): Promise<AssistantVersionPaginatedResponse>
```

--------------------------------

### API Client

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Import and use the pre-configured API client instance.

```typescript
import { client } from '@vapi-ai/web';

// Use the API client directly
await client.assistant.assistantControllerCreate({...});
```

--------------------------------

### listWorkflows()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Lists all workflows.

```typescript
client.workflow.workflowControllerFindAll(): Promise<Workflow[]>
```

--------------------------------

### Create Assistant

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Creates a new assistant using the API client.

```typescript
client.assistant.assistantControllerCreate(data: CreateAssistantDTO): Promise<Assistant>
```

--------------------------------

### Listen for Audio/Permissions Errors

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

This snippet shows how to specifically listen for errors related to accessing cameras or microphones.

```typescript
vapi.on('camera-error', (error) => {
  console.error('Camera/Audio Error:', error);
});
```

--------------------------------

### Call Recording

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/README.md

Starts a call with recording enabled and listens for the 'recording-stopped' event to log the recording URL.

```typescript
const call = await vapi.start('assistant-id', {
  recordingEnabled: true
});

vapi.on('recording-stopped', (event) => {
  console.log('Recording URL:', event.recordingUrl);
});
```

--------------------------------

### Event Types

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Demonstrates strongly-typed event handling using the event emitter with VapiClientToServerMessage.

```typescript
import type { VapiClientToServerMessage } from '@vapi-ai/web';

// Valid message types
type ValidMessages = 
  | AddMessageMessage
  | ControlMessages
  | SayMessage
  | EndCallMessage;
```

--------------------------------

### Variable Substitution

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/README.md

Starts a call with predefined variable values that can be used for substitution within the assistant's prompts.

```typescript
const call = await vapi.start('assistant-id', {
  variableValues: {
    'customer_name': 'Alice',
    'account_id': 'ACC-123',
    'balance': '$500'
  }
});
// If assistant says "Hello {{customer_name}}", 
// it becomes "Hello Alice"
```

--------------------------------

### Use Browser DevTools

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

Demonstrates how to use browser console methods like `console.table` to inspect Vapi events for debugging.

```typescript
// In browser console
vapi.on('error', (e) => console.table(e));
vapi.on('call-start-progress', (e) => console.table(e));
```

--------------------------------

### Source Code Organization

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/README.md

Directory structure of the client-sdk-web project.

```bash
client-sdk-web/
├── vapi.ts                 # Main Vapi class (1640 lines)
├── api.ts                  # Generated API client (46K+ lines)
├── client.ts               # API client setup (19 lines)
├── daily-guards.ts         # Safety utilities (82 lines)
├── __tests__/              # Test files
├── example/                # Example application
├── package.json            # Dependencies
└── tsconfig.json          # TypeScript config
```

--------------------------------

### Network Diagnostics

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/README.md

Runs network tests before starting a call to check for connectivity issues and displays an error message if the network is poor.

```typescript
// Before starting calls
const results = await Vapi.runNetworkTestsStandalone();
if (results.networkConnectivity.result === 'error') {
  showErrorMessage('Poor network connection detected');
} else {
  await vapi.start('assistant-id');
}
```

--------------------------------

### Chrome 140+ Compatibility Fixes

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

Illustrates how the Vapi SDK automatically corrects known issues for Chrome 140+ regarding `audioSource` and `alwaysIncludeMicInPermissionPrompt` configurations.

```typescript
// Auto-fixed: audioSource: false → audioSource: true
const vapi = new Vapi(
  'api-token',
  undefined,
  undefined,
  { audioSource: false } // Automatically corrected to true
);

// Auto-fixed: alwaysIncludeMicInPermissionPrompt: false → removed
const vapi = new Vapi(
  'api-token',
  undefined,
  { alwaysIncludeMicInPermissionPrompt: false } // Automatically handled
);
```

--------------------------------

### Create Web Call

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Creates a web call.

```typescript
client.call.callControllerCreateWebCall(data: CreateWebCallDTO): Promise<Call>
```

--------------------------------

### createOpenAiChat()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Creates a chat using OpenAI-compatible API.

```typescript
client.chat.chatControllerCreateOpenAiChat(data: OpenAIResponsesRequest): Promise<ResponseObject | ResponseTextDeltaEvent | ResponseTextDoneEvent | ResponseCompletedEvent | ResponseErrorEvent>
```

--------------------------------

### VideoPlan Configuration

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Configuration object for video-related settings, such as enabling video recording.

```typescript
videoPlan?: {
  videoRecordingEnabled?: boolean;
  // ... video-specific settings
}
```

--------------------------------

### createWorkflow()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Creates a new workflow.

```typescript
client.workflow.workflowControllerCreate(data: CreateWorkflowDTO): Promise<Workflow>
```

--------------------------------

### Handling Input Validation Errors

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

Demonstrates how to catch and handle specific input validation errors, such as attempting to start a call when one is already in progress or when required parameters are missing.

```typescript
try {
  await vapi.start(assistantId);
} catch (error) {
  if (error.message.includes('already started')) {
    // Stop current call first
    await vapi.stop();
    await vapi.start(assistantId);
  } else if (error.message.includes('must be provided')) {
    console.error('Valid assistant ID is required');
  }
}
```

--------------------------------

### Integration with Vapi Class

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/daily-guards.md

Illustrates how the Vapi class internally utilizes safety functions for Daily.co configuration and method calls.

```typescript
// In vapi.ts constructor
this.dailyCallConfig = createSafeDailyConfig(dailyCallConfig);
this.dailyCallObject = createSafeDailyFactoryOptions(dailyCallObject);

// During call setup
safeSetLocalAudio(this.call, !mute);

// For device changes
await safeSetInputDevicesAsync(this.call, options);
```

--------------------------------

### roomDeleteOnUserLeaveEnabled Example

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Controls room lifecycle when user leaves. `true`: Room is deleted when user leaves, preventing reconnection. `false`: Room persists, allowing reconnection via `reconnect()`

```typescript
// Allow reconnection
const call = await vapi.start(
  'assistant-id',
  undefined,
  undefined,
  undefined,
  undefined,
  { roomDeleteOnUserLeaveEnabled: false }
);

// Later, reconnect to the same call
await vapi.reconnect(call);
```

--------------------------------

### Partial Types for DTOs

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Demonstrates how to use the `Partial` utility type to allow partial updates for DTOs, specifically for updating an assistant.

```typescript
type UpdateAssistantDTO = Partial<CreateAssistantDTO>;
// Allows updating only some fields:
await client.assistant.assistantControllerUpdate('id', {
  name: 'New Name'
  // Don't need to provide model, voice, etc.
});
```

--------------------------------

### Enable Detailed Logging

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

Sets up listeners for various Vapi events and logs them with timestamps to an array. It also provides a function to export the log as a JSON string.

```typescript
// Track all events for debugging
const eventLog: Array<{ event: string; data: any; time: Date }> = [];

const events = [
  'call-start', 'call-end', 'speech-start', 'speech-end',
  'message', 'error', 'camera-error', 'call-start-progress',
  'call-start-failed', 'network-quality-change'
];

events.forEach(event => {
  vapi.on(event as any, (data: any) => {
    eventLog.push({ event, data, time: new Date() });
    console.log(`[${event}]`, data);
  });
});

// Export log for analysis
function exportLog() {
  return JSON.stringify(eventLog, null, 2);
}
```

--------------------------------

### Pattern 3: API-First

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/modules.md

Use the API client directly without the high-level Vapi class.

```typescript
import { client } from '@vapi-ai/web';

// Use API directly without the high-level Vapi class
const call = await client.call.callControllerCreate({
  assistantId: 'assistant-123',
  customerNumber: '+1234567890',
  phoneNumberId: 'phone-456'
});
```

--------------------------------

### SafeDailyFactoryOptions Interface

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/daily-guards.md

Type-safe wrapper for Daily.co factory options with audio source validation.

```typescript
interface SafeDailyFactoryOptions extends Omit<DailyFactoryOptions, 'audioSource'> {
  audioSource?: string | boolean | MediaStreamTrack;
}
```

--------------------------------

### createChat()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Creates a new chat conversation.

```typescript
client.chat.chatControllerCreateChat(data: CreateChatDTO): Promise<Chat | CreateChatStreamResponse>
```

--------------------------------

### Manual Release (Emergency)

Source: https://github.com/vapiai/client-sdk-web/blob/main/RELEASE.md

Commands to manually publish a release if the automated process fails.

```bash
npm ci
npm run build
npm test
npm publish --access public
```

--------------------------------

### List paginated assistants

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

List paginated assistants

```typescript
client.v2.assistantControllerFindAllPaginated(query?: {
  page?: number;          // Page number (min 1)
  sortOrder?: 'ASC' | 'DESC';  // Sort order
  limit?: number;         // Items per page (0-1000)
  // ... date filters
}): Promise<AssistantPaginatedResponse>
```

--------------------------------

### setOutputDeviceAsync() Method Signature

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/vapi-class.md

The signature for the setOutputDeviceAsync() method, used to asynchronously set the audio output device.

```typescript
setOutputDeviceAsync(options: Parameters<DailyCall['setOutputDeviceAsync']>[0]): void
```

--------------------------------

### List Assistants

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Lists all assistants with optional filtering and pagination.

```typescript
client.assistant.assistantControllerFindAll(query?: {
  limit?: number;
  createdAtGt?: string;
  createdAtLt?: string;
  createdAtGe?: string;
  createdAtLe?: string;
  updatedAtGt?: string;
  updatedAtLt?: string;
  updatedAtGe?: string;
  updatedAtLe?: string;
}): Promise<Assistant[]>
```

--------------------------------

### createSquad()

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/api-client.md

Creates a new squad (group of assistants).

```typescript
client.squad.squadControllerCreate(data: CreateSquadDTO): Promise<Squad>
```

--------------------------------

### Permission Handling

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/errors.md

Handles camera and microphone permission errors. It displays user-friendly messages when permissions are denied.

```typescript
vapi.on('camera-error', (error) => {
  console.error('Camera permission error:', error);
  
  // Show user-friendly message
  showDialog('Camera access denied. Please check your browser permissions.');
});

vapi.on('error', (error) => {
  if (error?.error?.message?.includes('NotAllowedError')) {
    showDialog('Microphone permission is required to use this feature.');
  }
});
```

--------------------------------

### Function Tool Configuration

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/configuration.md

Specifies the configuration for a function tool, including its name, description, parameters, and a server webhook URL for execution.

```typescript
{
  type: 'function',
  function: {
    name: 'get_weather',
    description: 'Get current weather',
    parameters: {
      type: 'object',
      properties: {
        location: {
          type: 'string',
          description: 'City name'
        }
      },
      required: ['location']
    }
  },
  server: {
    url: 'https://your-api.com/webhook',
    secret: 'webhook-secret'
  }
}
```

--------------------------------

### safeSetInputDevicesAsync Function Signature

Source: https://github.com/vapiai/client-sdk-web/blob/main/_autodocs/api-reference/daily-guards.md

Safely sets input devices (microphone/audio source) with validation to prevent Chrome 140+ issues.

```typescript
async function safeSetInputDevicesAsync(
  call: DailyCall | null,
  options: Parameters<DailyCall['setInputDevicesAsync']>[0]
): Promise<void>
```