Live ElevenLabs conversation
Streams microphone audio to an ElevenLabs agent, plays responses through Reachy’s speaker, and handles interruptions naturally.
Reachy Mini application · v0.3.0
Reachy Eleven connects ElevenLabs conversational AI to Reachy Mini, pairing natural voice interaction with camera awareness and expressive, coordinated movement.
01 ElevenLabs voice
02 Optional vision
03 Coordinated motion
04 Dashboard setup
What it does
One app owns the conversation loop and coordinates every movement source so speech, tracking, and model-triggered gestures work together.
Streams microphone audio to an ElevenLabs agent, plays responses through Reachy’s speaker, and handles interruptions naturally.
Answers questions about the current frame through an OpenAI or OpenAI-compatible vision endpoint.
Coordinates head pose, antennas, body yaw, expressions, speech motion, sleep, wake, and optional visual tracking.
Save agent settings, start or stop conversations, inspect runtime state, and recover from transient daemon disconnects.
How it works
The app runs on Reachy Mini, uses the daemon's media stream, and keeps voice, camera tools, and motion under one lifecycle.
Start a conversation
uv sync --dev
uv run reachy-eleven-vision
Local development · Python 3.11+