syxanash/maxheadbox

What it solves

Max Headbox provides a way to create a fully local, voice-activated AI assistant that lives on a Raspberry Pi. It removes the need for cloud-based AI providers, ensuring privacy and offline functionality while allowing the AI to interact with and control physical hardware on the device.

How it works

The system uses a combination of local tools for voice processing and LLM inference:

• Voice Activation: Uses Vosk for wake-word detection and faster-whisper for transcribing speech to text.
• Brain: Runs LLMs locally via Ollama (specifically using models like gemma3:1b and gemma4:e2b).
• Execution: A frontend built with Node.js/Express handles the agent logic and tool orchestration, while a Python backend manages the Raspberry Pi's hardware interactions.
• Tooling: Users can define custom JavaScript modules to allow the agent to perform specific actions, with a "dangerous" flag available to require manual user confirmation for sensitive tasks.

Who it’s for

It is designed for hobbyists and makers who want to build a physical AI desk toy or a local home automation hub using a Raspberry Pi 5.

Highlights

• 100% Local: Operates entirely on-device without cloud dependencies.
• Voice-Driven: Integrated wake-word and transcription pipeline.
• Extensible Tooling: Simple JS-based system for adding custom automation capabilities.
• Safety Guardrails: Built-in confirmation flow for tools marked as dangerous.
• Visual Feedback: Includes an animated character UI to indicate the agent's state (thinking, sleeping, etc.).