@atrix.dev/ai-voice-sdk v1.0.7

AI Voice Caller SDK

Create AI-powered voice calls with natural language processing and voice synthesis.

Features

Installation

npm install @atrix.dev/ai-voice-sdk

Quick Start

import { AIVoiceCaller } from '@atrix.dev/ai-voice-sdk';

// Initialize the SDK
const voiceCaller = new AIVoiceCaller({
  apiKey: 'your_api_key',
  twilioAccountSid: 'your_twilio_sid',
  twilioAuthToken: 'your_twilio_token',
  twilioPhoneNumber: 'your_twilio_number',
  elevenLabsApiKey: 'your_elevenlabs_key',
  elevenLabsVoiceId: 'your_voice_id'
});

// Make a call
const result = await voiceCaller.makeCall({
  to: '+1234567890',
  message: 'Hello from AI Voice!'
});

// Send an SMS
const sms = await voiceCaller.sendSMS(
  '+1234567890', 
  'Hello from AI Voice!'
);

Configuration Options

API Reference

Voice Calls

// Make a voice call
const call = await voiceCaller.makeCall({
  to: '+1234567890',
  message: 'Hello!',
  voiceSettings: {
    stability: 0.7,
    similarityBoost: 0.5,
    style: 0.5,
    useSpeakerBoost: true
  }
});

// End a call
await voiceCaller.endCall('call_id');

// Get call history
const history = await voiceCaller.getCallHistory(10);

// Get call metrics
const metrics = await voiceCaller.getCallMetrics('call_id');

SMS Messaging

// Send SMS
const sms = await voiceCaller.sendSMS(
  '+1234567890',
  'Hello from AI Voice!'
);

Real-time Monitoring

// Listen for call status updates
voiceCaller.onCallStatus((status) => {
  console.log('Call status:', status);
});

// Listen for errors
voiceCaller.onError((error) => {
  console.error('Error:', error);
});

Analytics & Metrics

// Get aggregate metrics
const metrics = await voiceCaller.getMetrics();

// Get conversation transcript
const transcript = await voiceCaller.getConversationTranscript('call_id', {
  format: 'json',
  includeTiming: true,
  speakerLabels: true
});

Error Handling

Best Practices

1. Rate Limiting

   • Monitor your API usage
   • Implement proper error handling
   • Use the built-in rate limiter

2. Resource Management

   • Clean up resources with disconnect()
   • Monitor memory usage with metrics
   • Use caching for frequent requests

3. Error Handling

   • Always implement error handlers
   • Use try-catch blocks
   • Monitor error rates

4. Performance

   • Enable caching for repeated calls
   • Use concurrent call limits
   • Monitor call metrics

Examples

Advanced Call Setup

const call = await voiceCaller.makeCall({
  to: '+1234567890',
  message: 'Hello!',
  voiceSettings: {
    stability: 0.7,
    similarityBoost: 0.5,
    style: 0.5,
    useSpeakerBoost: true
  },
  webhookEvents: ['initiated', 'ringing', 'answered', 'completed'],
  callbackUrl: 'https://your-webhook.com/callback'
});

Health Monitoring

// Check service health
const isHealthy = await voiceCaller.checkHealth();

// Clear cache if needed
await voiceCaller.clearCache();

// Disconnect services
voiceCaller.disconnect();

Support

• Documentation: https://docs.atrix.dev/developers/ai-voice-sdk
• Issues: GitHub Issues
• Email: support@atrix.dev

License

MIT