Hello World One App to Rule Them All

The most complete web dashboard for Pollen Robotics' Reachy Mini. 146 APIs, 31 AI tools, computer vision, 81 emotions, MuJoCo 3D simulation, voice pipeline, and full system telemetry — all from your browser.

146
API Endpoints
81
Emotions
31
AI Tools
20
Dances
20
3D Scenes
21
Robot Skins
Try Live Demo → Read the Full Documentation →

Full dashboard walkthrough — every feature in under 2 minutes

System Monitoring

Real-Time System Telemetry

Complete visibility into your robot's hardware state with live-updating charts and diagnostics.

Status Dashboard

Status Dashboard

Nine real-time charts streaming over WebSocket at configurable rates. See your robot's health at a glance.

  • CPU temperature, load average, and per-core utilization
  • RAM breakdown (Used, Buffer, Cache) and disk I/O
  • Network throughput and WiFi signal strength
  • Fan RPM, thread count (App vs Daemon)
  • Top processes with CPU/memory usage
  • Full hardware inventory (board, serial, kernel, firmware)
WebSocket 15Hz Robot / 1Hz Stats Raspberry Pi CM4
Head Telemetry

Head Telemetry

Live streaming charts for every degree of freedom in Reachy's Stewart platform head mechanism.

  • Roll, Pitch, Yaw rotation with real-time graphing
  • X, Y, Z translation position tracking
  • Left and Right antenna angle monitoring
  • Stewart Platform actuator visualization (6 motors)
  • Head Pose summary with numeric readout
6-DOF Stewart Platform 9 Dynamixel Servos
Control & Visualization

Floating Panels & 3D Simulation

Draggable, resizable floating panels let you compose your ideal workspace.

Floating Panels

Composable Floating Panels

Two persistent panels hover above every tab, creating a fully customisable workspace.

  • Camera Panel — Live WebRTC feed, snapshot, video/audio recording, listen/speak
  • Joystick Panel — 5-axis head control with return-to-center toggle
  • Supports robot camera, laptop webcam, iPhone Continuity Camera
  • Draggable, resizable, minimisable — position remembered across sessions
Draggable Resizable Multi-Camera
MuJoCo 3D Simulation

MuJoCo 3D Simulation

Real-time 3D visualization of Reachy Mini powered by MuJoCo and Three.js. The robot's pose mirrors the physical hardware in real time with smooth-shaded GLB meshes.

MuJoCo Three.js Real-time Sync
Joystick & Cameras

Joystick & Multi-Camera

Intuitive joystick controls for head movement (L/XY/R), look direction, Z/Roll, body rotation, and a reset button. Multiple camera feeds in floating panels — robot camera, FaceTime, and iPhone via WebRTC.

6-axis Control WebRTC Cameras Return to Center
Camera source selector showing iPhone and FaceTime cameras

Any Camera, Anywhere

Use your iPhone as a wireless camera via Apple Continuity Camera, your MacBook's FaceTime HD camera, or Reachy's onboard camera — all selectable from a single dropdown. Every video source works with YOLO vision, Follow mode, and recording.

  • Reachy's Camera — onboard wide-angle via CSI-2
  • iPhone Continuity Camera — wireless, zero-config
  • FaceTime HD — laptop webcam via WebRTC
  • Hot-swap between sources without restarting
Continuity Camera WebRTC Hot-Swap
Chrome flags setting for insecure origins treated as secure

One-Time Browser Setup

Since the dashboard runs on HTTP (not HTTPS), Chrome blocks camera access by default. A one-time flag enables WebRTC on your local network — paste your dashboard URL into chrome://flags/#unsafely-treat-insecure-origin-as-secure and set it to Enabled.

  • Navigate to chrome://flags and search for "insecure"
  • Add your dashboard URL (e.g. http://reachy-mini.local:8042)
  • Set to Enabled and relaunch Chrome
  • iPhone, FaceTime, and laptop cameras now appear in the dropdown
Required for Camera chrome://flags One-Time Setup
Expressive Movement

81 Emotions, 20 Dances, 6 Oscillations

A rich library of pre-built movements organized by mood category, with live preview in the 3D simulation.

Emotions Gallery

Emotions Gallery

Browse and trigger 81 unique emotional expressions, filterable by mood categories.

  • Categories: Angry, Confused, Curious, Dance, Fear, Happy, Mood, Response
  • Each emotion card shows name, duration, and description
  • Auto-fit toggle adapts motion to robot's range
  • Real-time preview in 3D simulation and on physical robot
  • Search bar for quick access to specific emotions
81 Emotions 20 Dances HuggingFace Datasets
Oscillation Patterns

Oscillation Patterns

Continuous, looping head movement patterns with adjustable amplitude and speed for natural idle behaviour.

  • Sway — Gentle side-to-side head movement
  • Nod — Slow vertical nodding
  • Circle — Circular head motion
  • Breathe — Subtle breathing-like oscillation
  • Figure 8 — Figure-eight head pattern
  • Look Around — Curious, exploratory movement
Amplitude Control Speed Control
AI-Powered

Conversational AI with Tool Use

Chat with Reachy using multi-provider LLMs. The AI can express emotions, set timers, create content, and control the robot.

AI Conversation

Multi-Provider AI Chat

Full conversation interface with timestamped message log showing confidence and volume. Powered by LiteLLM for unified access to Anthropic, OpenAI, Ollama, and more. The AI triggers emotional expressions automatically based on context.

Claude OpenAI Ollama LiteLLM
AI Tool Use

31 Built-in AI Tools

The LLM can call robot-native tools: set timers, trigger emotions, create scratchpad content, play music, control lights, take snapshots, detect objects, and more. Tool calls are displayed inline in the conversation with full transparency.

set_timer create_scratchpad play_emotion 31 Tools Total
Generative UI

Scratchpad — Rich Content from Conversation

The AI creates rich HTML content on demand: recipes, diagrams, charts, and more — all rendered beautifully in the Scratchpad tab.

Scratchpad Recipe

AI-Generated Content

Ask Reachy for a pancake recipe and it creates a beautifully formatted page with ingredients, step-by-step instructions, and tips — complete with a themed robot illustration.

Markdown Rendering Theme-Aware
Scratchpad Diagrams

Mermaid Diagrams & Charts

Interactive architecture diagrams rendered with Mermaid.js, pie charts, activity heatmaps, and more. The AI can generate visual documentation of the robot's own architecture.

Mermaid.js Charts Searchable
Smart Assistant

Timers, Alarms & Ambient Sounds

Set timers via conversation or the UI. Built-in ambient sound library for focus, relaxation, or atmosphere.

Timers and Ambient Sounds

Timers, Alarms & Soundscapes

A complete timer and alarm system with rich sound options.

  • Create timers via AI conversation ("Set a timer for 5 mins")
  • Named alarms with one-shot or recurring schedules
  • 9 ambient sounds: Rain, Ocean, Forest, Crickets, Fireplace, Wind, White/Pink/Brown Noise
  • 3 timer sounds: Bell, Ding, Success
  • 6 alarm sounds: Birds, Classic Bell, Digital, Gentle Chime, Harp, Marimba
  • Upload custom sounds with Add Sound
Voice-Activated 9 Ambient Sounds Custom Upload
Computer Vision

YOLO Vision — Real-Time Detection

On-device and in-browser computer vision with pose estimation, object detection, and segmentation.

Vision Settings

YOLO Vision Configuration

Three processing modes: Off, CM4 (on-robot), and WebGPU (in-browser via WASM). Choose from pose keypoints, detection, or segmentation tasks. Adjustable FPS, confidence threshold, and model size selection.

WebGPU / WASM CM4 On-Device YOLO Nano
Pose Detection

Real-Time Pose Detection

Blue skeleton overlay tracks body keypoints in real time from any connected camera. Supports Follow mode where the robot tracks detected poses. Works with robot camera, FaceTime, or iPhone sources.

Pose Keypoints Follow Mode Multi-Camera Source
Developer Tools

In-Browser Shell Terminal

Full bash terminal access via WebSocket PTY with shared tmux sessions.

Shell Terminal

WebSocket Terminal

A fully interactive terminal embedded in the dashboard, powered by tmux for persistent, shared sessions.

  • Bidirectional PTY via WebSocket
  • Terminal and Tmux tab modes
  • Shared sessions — multiple users, same terminal
  • Connection status indicator
  • Full process visibility (ps aux)
WebSocket PTY tmux Shared Sessions
Media Management

Snapshots, Recordings & Music

Capture, record, and playback media directly from the dashboard with thumbnail previews and metadata.

Media Library

Media Library

A built-in media manager for all your robot's captures and audio content.

  • Camera snapshots with preview grid
  • Video recordings with duration and thumbnails
  • Audio recordings with waveform visualization
  • Music library with ID3 metadata and cover art
  • Upload, download, and delete with drag-and-drop
  • NVR (Network Video Recorder) camera snapshots
Snapshots Video Recording Music Playback
Audio & Connectivity

Volume Mixer, Audio Routing & Bluetooth

Fine-grained audio control with channel mixing and Bluetooth A2DP device management.

Audio Settings

Audio Routing & Volume Mixer

Complete audio pipeline management with per-channel volume control.

  • Mic Input and Speaker routing (Robot Mic / Robot Speaker)
  • 6 volume channels: Hardware, Master, Speech, Music, Effects, Ambient
  • Bluetooth device scanning and pairing
  • A2DP (high-quality music) and HSP (voice) profiles
  • AuraBox LED speaker integration
  • Test button for audio output verification
6 Volume Channels Bluetooth A2DP GStreamer Pipeline
Configurability

Tuning & LLM Behaviour

Full control over voice recognition thresholds and the LLM's personality through an editable system prompt.

Tuning and LLM Settings

Tuning & System Prompt

Adjust voice recognition sensitivity and customise Reachy's AI personality.

  • Confidence threshold for speech recognition
  • Volume threshold to filter background noise
  • Microphone gain control
  • Editable system prompt defining Reachy's personality
  • TTS speech output rules built into the prompt
  • Save/Reset prompt functionality
Whisper STT Kokoro TTS Editable Prompt
Customisation

20 Scenes & 21 Robot Skins

Personalise the 3D simulation with procedural environments and paint Reachy in country flags, metallic finishes, or vibrant gradients.

3D Scenes and Robot Skins

Scenes & Skins Gallery

A rich collection of 3D environments and robot textures, instantly switchable from the Appearance settings.

  • 20 scenes: Default, Studio, Sunset, Ocean, Night Sky, Forest, Neon, Rainy City, Winter, Desert, Underwater Cave, Mars, Crystal Cave, Volcano, Candy Land, Zen Garden, Haunted, Space Station, Clouds
  • 21 skins: Default, France, Japan, USA, Brazil, Germany, Italy, UK, India, Australia, Gold, Chrome, Matte Black, Carbon, Rainbow, Neon, Sunset, Ocean, Forest, Pollen, Robot
20 Scenes 21 Skins Instant Switch
Crystal Cave Scene

Crystal Cave

Clouds Scene

Clouds

Brazil Skin

Brazil Skin

Architecture

Built With

Pure Python + vanilla JavaScript. No framework, no bundler, no build step.

⚡

FastAPI

146 REST endpoints + 6 WebSockets

🌐

Vanilla JS

Zero dependencies, SPA

🤖

MuJoCo + Three.js

Physics-accurate 3D simulation

👁

YOLO + WebGPU

Real-time vision in browser

🗣

Whisper STT

Voice recognition pipeline

🔈

Kokoro TTS

Text-to-speech engine

🧠

LiteLLM

Multi-provider AI (Claude, GPT, Ollama)

🔌

WebSocket-First

All real-time data, zero polling