First! The first Reachy Mini app built for the new version of Reachy Mini, by Clem Delangue, co-founder & CEO of Hugging Face.
Reachy Mini · Python app

☀️ Good Morning

Press a button and Reachy Mini greets you out loud.

↑ Preview only. The real button lives at localhost:7860 when the app is running on your robot.

Get a Reachy Mini

How to use it on your robot

  1. 1

    Open Discover Apps

    On your Reachy Mini's dashboard, click Discover Apps.

  2. 2

    Find "Good Morning"

    Search for good-morning or clem. Click Install.

  3. 3

    Start the app

    Hit Start on the good_morning tile. It boots a tiny web server on the robot.

  4. 4

    Open the button

    Once the app is running, go to http://localhost:7860 in your browser. The app also auto-opens this page on launch. Click the orange button → Reachy Mini speaks.

No robot? No problem.

Try it in simulation

The Reachy Mini desktop app ships with a built-in MuJoCo simulator — apps, audio, and motion all work just like with a real robot. Install Good Morning from the dashboard and watch a simulated Reachy Mini wobble and speak.

  1. 1

    Install the Reachy Mini desktop app

    Download it from the Reachy Mini desktop app repo for your OS (macOS, Linux, Windows). Open it — it bundles the daemon and the simulator, so there's nothing else to install.

  2. 2

    Switch to simulation mode

    In the desktop app, choose simulation as the connection mode (instead of USB or wireless). The 3D viewport shows a simulated Reachy Mini you can drive exactly like the physical one.

  3. 3

    Install & start Good Morning

    Discover Apps → search good-morning → Install → Start. Your browser opens at localhost:7860; click the button and the simulated robot will speak and move just like the real one.

More details: Reachy Mini Simulation guide · SDK on GitHub

How it works

  1. 1

    Tap the button

    Click the orange button at localhost:7860 in your browser.

  2. 2

    POST /say

    The page sends a one-line POST to the FastAPI server running inside the app.

  3. 3

    Synthesised once

    The phrase is rendered with offline TTS (pyttsx3) and cached as a WAV.

  4. 4

    Played on the robot

    Audio is resampled to 16 kHz mono float32 and pushed via media.push_audio_sample.

Features