Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Reachy Mini MCP wiggle your antennas"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP (Model Context Protocol) server for controlling Reachy Mini robot from Claude.
uv syncRelated MCP server: Reachy Mini MCP
Add to ~/.claude/claude_code_config.json:
{
"mcpServers": {
"reachy-mini": {
"command": "uv",
"args": ["run", "--directory", "/path/to/reachy-mini-mcp", "reachy-mini-mcp"],
"env": {
"REACHY_HOST": "localhost"
}
}
}
}REACHY_HOST - Daemon host (default: localhost)
REACHY_PORT - Daemon port (default: 8000)
REACHY_SSH_USER - SSH username for streaming commands (default: reachy)
Tool | Description |
| Get daemon health status |
| Get full robot state |
| Move head (pitch, roll, yaw) |
| Look at world coordinates |
| Rotate body yaw |
| Play animation (wave, nod, dance) |
| Happy antenna expression |
| Sad antenna expression |
| Wiggle antennas |
| Enable motors (stiff) |
| Disable motors (limp) |
| Enable soft mode |
| Capture camera frame |
| Text-to-speech |
| Play sound file |
| Start WebRTC video/audio streaming |
| Stop WebRTC streaming |
| Check if stream is running |
Reachy Mini daemon running on port 8000
Python 3.10+
Capture a single frame from the robot's camera
Inputs: no named parameters.
Check if the robot's camera is working by capturing a test frame
Inputs: no named parameters.
Check if the robot's microphone is working by recording a short sample
Inputs: duration.
Get detailed motor status information
Inputs: no named parameters.
Get full robot state including joint positions and sensor data
Inputs: no named parameters.
Put the robot to sleep (execute sleep behavior)
Inputs: no named parameters.
Move the robot to its zero/neutral position
Inputs: no named parameters.
List all saved audio recordings on the robot
Inputs: no named parameters.
Move the robot's antennas to specified positions
Inputs: left, right, duration.
Move the robot's head to a specific orientation. Pitch/roll limited to [-40, 40] degrees, yaw to [-180, 180] degrees.
Inputs: yaw, roll, pitch, duration.
Play an audio recording through the robot's speaker
Inputs: filename (required).
Try to reconnect to the robot daemon
Inputs: no named parameters.
Start continuous sinusoidal movement on the robot
Inputs: no named parameters.
Start WebRTC video/audio streaming from the robot's camera and microphone. Returns the stream URL to connect to.
Inputs: no named parameters.
Get Reachy Mini daemon health status and robot connection info
Inputs: no named parameters.
Stop continuous movement on the robot
Inputs: no named parameters.
Stop the WebRTC video/audio stream on the robot
Inputs: no named parameters.
Wake up the robot (execute wake up behavior)
Inputs: no named parameters.
Check if WebRTC stream is currently running and get the stream URL
Inputs: no named parameters.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 19 tools
Most tools target clearly distinct actions such as moving the head, capturing images, or checking status. Some overlap exists among reachy_status, reachy_get_state, and reachy_get_motor_status, and between reachy_check_camera and reachy_capture_image, but the descriptions mostly clarify their different purposes.
All tools follow a consistent reachy_ prefix with verb_noun naming, such as move_head, capture_image, and stop_webrtc_stream. The pattern is uniform and predictable across the entire set.
At 19 tools, the count is slightly above the ideal sweet spot but remains justifiable for a robot control server covering movement, sensors, media, and streaming. The tools are not redundant enough to warrant trimming, though a few could potentially be consolidated.
The tool set covers the main robot lifecycle: wake/sleep, status, movement, camera, microphone, audio playback, and WebRTC streaming. Minor gaps exist such as no explicit individual joint control or audio file upload capability, but these do not critically undermine the server's apparent purpose.
Zero-auth ROS 2 Humanoid, Q3 Proxemics, 3-Level Voice Arbiter, A2A-MCP Bridge & W3C VC 2.0.
Operate Linux, macOS and Windows from your LLM. Every action runs through an auditable allowlist.
Drive real Android & iOS devices and web browsers from natural language for mobile + web QA. 290+ tools across device control, app management, automation sessions, browser automation, and flow recording / replay. Bearer-auth — get a token at robotactions.com → Profile → API Tokens.
Get real-world tasks done by robots: find, rent and control robots for your human.
