PiSugar/whisplay-ai-chatbot

What it solves

It provides the software and hardware blueprint for creating a pocket-sized, voice-interactive AI chatbot. It transforms a Raspberry Pi into a handheld device that can listen to speech, process it via an LLM, and talk back, functioning like a futuristic walkie-talkie.

How it works

The system runs on a Raspberry Pi (Zero 2w or Pi 5) equipped with a Whisplay HAT that provides an LCD screen, speaker, and microphone. It uses a combination of Python and Node.js to handle the chatbot logic. The project supports both online and offline builds, with optional integration for AI accelerator cards like the LLM8850 or Raspberry Pi AI HAT+ 2 (Hailo-10H) to speed up whisper (speech-to-text), LLM, and VLM (vision-language model) processing.

Who it’s for

Hardware enthusiasts, makers, and AI hobbyists who want to build a physical, portable AI assistant using Raspberry Pi hardware.

Highlights

• Voice-First Interaction: Supports push-to-talk and wake-word functionality for hands-free use.
• Multimodal Capabilities: Supports image generation from text prompts and vision-language models (VLM).
• Hardware Acceleration: Compatible with AI accelerator cards for faster local inference.
• Portable Design: Designed for use with PiSugar batteries and custom 3D-printable enclosures.