title: Reachy Nova emoji: "\U0001F916" colorFrom: indigo colorTo: cyan sdk: static pinned: false short_description: Voice, Vision & Browser AI for Reachy Mini tags:

• reachy_mini
• reachy_mini_python_app

Reachy Nova

AI-powered brain for Reachy Mini, integrating three Amazon Nova services:

• Nova Sonic - Real-time speech-to-speech conversation
• Nova 2 Lite - Camera-based vision and scene understanding
• Nova Act - Voice-controlled browser automation

This project enables Reachy Mini to see, hear, speak, and interact with the digital world.

Documentation

• Architecture Overview: High-level system design.
• Component Documentation: Detailed docs for each module.
  • Nova Sonic (Voice)
  • Nova Vision (Sight)
  • Nova Browser (Action)
  • Tracking Manager
  • Skills System
• Main Application: Core loop and state management.
• Setup Guide: detailed installation instructions.

Quick Start

1. Install dependencies:

   uv sync
   

2. Configure environment:

   cp .env.sample .env
   # Edit .env with your AWS credentials
   

3. Run:

   uv run python -m reachy_nova.main
   

Features

• Natural Conversation: Talk to the robot naturally - it responds with voice via Nova Sonic.
• Visual Understanding: The robot sees through its camera and describes what it sees.
• Browser Automation: Say "search for..." or "look up..." to trigger Google searches and more.
• Smart Tracking:
  • Tracks faces automatically.
  • Turns to look at speakers.
  • Reacts instantly to snaps/claps.
• Expressive Animations: Mood-reactive antenna and head movements.
• Web Dashboard: Monitor status at http://localhost:8042.

Project Structure

• reachy_nova/: Main package.
  • main.py: Entry point and orchestration.
  • nova_sonic.py: Voice interface.
  • nova_vision.py: Vision interface.
  • nova_browser.py: Browser automation.
  • tracking.py: Head tracking logic.
• tools/: Utility scripts (e.g., demos).
• docs/: Detailed documentation.