This project delivers a pocket-sized AI chatbot, transforming a Raspberry Pi Zero 2w or RPi 5 into a voice-interactive device. It targets DIY enthusiasts and users seeking a portable, futuristic communication tool, offering both online API integration and robust offline capabilities.

How It Works

The system leverages a Raspberry Pi paired with the PiSugar Whisplay HAT, which integrates an LCD, microphone, and speaker. It processes voice input, interacts with AI models (local or cloud-based) for responses, and outputs synthesized speech. A key design choice is enabling full offline operation using accelerators like the LLM8850, providing local ASR, TTS, and LLM inference.

Quick Start & Requirements

Installation involves cloning the repository, running bash install_dependencies.sh, bash build.sh, and bash run_chatbot.sh. For boot-on-start, sudo bash startup.sh is used.

• Hardware: Raspberry Pi Zero 2w or RPi 5 (8GB RAM recommended for offline builds). PiSugar Whisplay HAT. PiSugar 3 (1200mAh) battery. Optional: LLM8850 AI Accelerator for offline functionality.
• Software: Requires audio drivers for the Whisplay HAT.
• Links:
  • Test Video Playlist: https://www.youtube.com/watch?v=lOVA0Gui-4Q
  • Tutorial: https://www.youtube.com/watch?v=Nwu2DruSuyI
  • Offline RPi 5 Tutorial: https://www.youtube.com/watch?v=kFmhSTh167U
  • Pre-build Images: https://github.com/PiSugar/whisplay-ai-chatbot/wiki

Highlighted Details

• Image Generation: Supports OpenAI, Gemini, and Volcengine backends, displaying generated images on the device.
• Offline Mode: Full local ASR (Whisper/Vosk), TTS (Piper), and LLM capabilities are achievable, especially with LLM8850 integration.
• Hardware Integration: The PiSugar Whisplay HAT provides a compact, all-in-one interface with screen, mic, and speaker.
• Power Management: Integrates with pisugar-power-manager for on-screen battery level display, installable via wget https://cdn.pisugar.com/release/pisugar-power-manager.sh; bash pisugar-power-manager.sh -c release.

Maintenance & Community

The provided README does not detail specific contributors, community channels (e.g., Discord, Slack), or a public roadmap.

Licensing & Compatibility

The project is licensed under GPL-3.0. This copyleft license requires derivative works to be distributed under the same license, potentially impacting integration with proprietary software.

Limitations & Caveats

Offline builds, particularly on RPi 5, demand significant RAM (8GB recommended). The LLM8850 Qwen3 LLM API is explicitly noted as not supporting tools. Full offline functionality is heavily dependent on acquiring and configuring additional hardware like the LLM8850.