pollen-robotics/reachy_mini

Reachy Mini's SDK

Reachy Mini – Open‑source expressive robot

What it is

  • A small, 3‑DoF robot kit (wireless or USB‑connected) that you can assemble in a few hours.
  • Comes with a Python and JavaScript SDK so you can program its motions, camera, and speakers.
  • Designed for hobbyists, makers, and AI developers who want a physical platform to test LLM‑driven or other AI behaviours.

Key components

  • Hardware – Raspberry Pi CM4‑based board (wireless version) or a simple USB‑powered board (Lite). Includes servos for head, arms, and a microphone/speaker.
  • SDK – Python package (reachy_mini) for low‑level control (joint angles, trajectories) and higher‑level utilities (head pose helpers). A JavaScript SDK lets you drive the robot from a browser via WebRTC, enabling zero‑install web apps.
  • App ecosystem – Apps are hosted on Hugging Face Spaces and installed through the robot’s control UI. Example apps: a conversational agent (LLM‑powered), a radio player, and a hand‑tracking controller.
  • Simulation – MuJoCo‑based virtual robot so you can develop and test code without hardware.

Getting started

  1. Pick a platform – Wireless (RPi CM4 + battery), Lite (USB), or Simulation (MuJoCo). Links in the README lead to step‑by‑step guides.
  2. Install the SDK – Recommended to use uv (fast installer) or fall back to pip.
  3. Run a quick demo – Python example shows how to move the head in a few lines:
    from reachy_mini import ReachyMini
from reachy_mini.utils import create_head_pose

with ReachyMini() as mini:
    mini.goto_target(
        head=create_head_pose(z=10, roll=15, degrees=True, mm=True),
        duration=1.0,
    )
  4. Explore apps – Browse the Hugging Face app store, install the Conversation app to talk to the robot, or build your own web app using the JavaScript SDK.

Typical use cases

  • AI‑driven interaction – Connect LLMs (Claude, Codex, etc.) to make the robot converse, answer questions, or follow spoken commands.
  • Computer‑vision demos – Use the built‑in camera with hand‑tracking or object‑detection models to make the robot react to gestures.
  • Education & research – Prototype embodied AI behaviours, reinforcement‑learning agents, or human‑robot interaction studies on a cheap, open‑source platform.
  • Prototyping web‑based robot interfaces – Build a browser UI that streams video and sends control commands via WebRTC, then share the URL with anyone.

Community & support

  • Discord – Active community for troubleshooting, sharing builds, and collaborating on apps.
  • Documentation – Hosted on Hugging Face; includes installation, quickstart, core concepts, and AI integration guides.
  • Contributing – Open issues are welcomed; contribution guidelines are in docs/contributing.md.

License

  • Code: Apache 2.0
  • Hardware design files: Creative Commons BY‑SA‑NC

Reachy Mini bridges the gap between software‑only AI projects and real‑world embodied agents, giving developers a tangible playground for LLM‑driven robotics.

Related

  • Dispatch
    Reachy Mini: Open-Source Desktop Robot for AI Development

    Hugging Face and Pollen Robotics have introduced Reachy Mini, an open-source, programmable desktop robot starting at $399 designed for human-robot interaction and AI experimentation.

  • Project
    pollen-robotics/reachy_mini_conversation_app

    A Python app that gives the Reachy Mini robot real‑time voice conversation, vision, and motion by connecting to Hugging Face realtime speech services and an LLM. It supports local VAD, persistent memory, extensible tool plugins (dance, head tracking, web search, weather, etc.), and a web UI for managing personalities and settings.

  • Dispatch
    Reachy Mini Local Speech Backend Integration

    Hugging Face has released a local speech-to-speech pipeline for Reachy Mini, allowing the robot to handle conversations fully locally using a cascaded VAD, STT, LLM, and TTS stack.

  • Dispatch
    Hugging Face acquires Pollen Robotics to expand open-source robotics hardware

    Hugging Face has acquired Pollen Robotics to integrate open-source humanoid hardware with the LeRobot software ecosystem, starting with the sale of the Reachy 2 robot.

  • Dispatch
    Hugging Face Open Responses Inference Standard

    Hugging Face has introduced Open Responses, an open inference standard based on the Responses API designed to replace the Chat Completion format for agentic AI workflows.