Skip to main content
Pipecat is a framework for voice agents — it wires together speech-to-text, an LLM, and text-to-speech into one pipeline with interruption handling. The Maya integration is here:

pipecat-ai/pipecat · PR #5222

The Maya TTS service for Pipecat.
Pipecat drives the WebSocket, so everything on the voice agents page applies — one socket per session, a context_id per turn, and barge-in by cancel. The integration handles those for you.