ClawdTalk

Multi-channel AI agent platform with real-time voice powered by OpenAI Realtime API.

Features

• 📞 Voice Calls - Natural, low-latency conversations with OpenAI Realtime
• 💬 SMS - AI-powered text messaging
• 🎙️ Real-time Streaming - Bidirectional audio via Telnyx WebSocket
• 🔊 Server-side VAD - Intelligent turn detection

Architecture

Phone Call → Telnyx → WebSocket Media Stream → OpenAI Realtime API
                ↓                                    ↓
           PCMU Audio ←←←←←←←←←←←←←←←←←←←←←←← AI Response

Quick Start

1. Install dependencies

npm install

2. Configure environment

cp .env.example .env
# Edit .env with your API keys

Required keys:

• TELNYX_API_KEY - From Telnyx Portal
• OPENAI_API_KEY - From OpenAI

3. Start the server

npm run dev

4. Expose with ngrok

ngrok http 3000

5. Configure Telnyx

1. Go to Voice → TeXML Applications in Telnyx Portal
2. Create a new application
3. Set webhook URL: https://your-ngrok-url.ngrok.app/api/webhook/voice
4. Assign your phone number

Voice Configuration

Available voices (set via OPENAI_VOICE):

• alloy (default)
• echo
• shimmer
• ash
• ballad
• coral
• sage
• verse

Development

# Run with hot reload (Next.js only, no WebSocket)
npm run dev:next

# Run with WebSocket support
npm run dev

# Production build
npm run build
npm start

How It Works

1. Incoming Call: Telnyx sends webhook to /api/webhook/voice
2. Answer: Server answers with streaming configuration
3. Media Stream: Telnyx opens WebSocket to /media-stream
4. Audio Bridge: Server connects to OpenAI Realtime API
5. Real-time Conversation: Audio flows bidirectionally with sub-second latency

License

MIT