@plaios/voice-live
v0.1.8
Published
Live voice input: OpenAI Realtime (WebRTC) agent that speaks in real time and calls every PLAIOS capability as a tool
Maintainers
Readme
Live Voice input
@plaios/voice-live adds a real-time voice agent to the PLAIOS web UI.
The browser opens a WebRTC call to the OpenAI Realtime API; the backend only
mints short-lived client secrets and executes tools, so the API key never
reaches the browser.
Every capability registered in discovery is exposed to the voice model as a
function tool. When the model calls one, the microfrontend forwards it to the
runtime (callTool), which runs the capability and returns the result to the
model, which then answers out loud.
Config: OPENAI_API_KEY (required), VOICE_LIVE_MODEL (default gpt-realtime),
VOICE_LIVE_VOICE (default marin), VOICE_LIVE_INSTRUCTIONS.
