fastrtc

Why we included this project

FastRTC turns an ordinary Python function into a live audio or video stream, so a voice assistant or video feature behaves like a conversation instead of a request-response round trip. The library handles the WebRTC and WebSocket transport, which saves you from writing the signaling and media plumbing yourself. It also includes voice-activity detection and turn taking, so pause detection and reply timing work out of the box, and optional speech-to-text and text-to-speech extras complete a spoken dialog loop. Since the same stream can mount onto a FastAPI app or launch through a built-in Gradio UI, you can move from a prototype to an endpoint inside your existing backend without switching stacks.