# Voice Agent API

AssemblyAI's Voice Agent API is a production speech-to-speech stack for building real-time voice agents over a single WebSocket.

Canonical URL: https://www.assemblyai.com/products/voice-agent-api

## Summary

The Voice Agent API handles speech recognition, turn detection, LLM reasoning, voice generation, interruptions, session handling, and tool calling. It is designed for teams that want the fastest path to a working production voice agent without stitching together separate STT, LLM, TTS, and orchestration providers.

## Key Capabilities

- Single WebSocket for audio input and audio output.
- Built on AssemblyAI's real-time speech recognition.
- Natural turn detection and interruption handling.
- Tool calling for application actions.
- Unified billing at $4.50/hr ($0.075/min).
- PCI-certified end-to-end.

## Common Use Cases

- Customer support voice agents.
- Front-office scheduling and routing.
- Healthcare and medical intake agents.
- Sales qualification agents.
- Field service and logistics agents.

## Related References

- https://www.assemblyai.com/docs/voice-agents/voice-agent-api.md
- https://www.assemblyai.com/docs/streaming/evaluations/voice-agents.md
- https://www.assemblyai.com/docs/integrations/twilio.md
- https://www.assemblyai.com/docs/integrations/telnyx.md
- https://www.assemblyai.com/pricing.md