PiSugar/whisplay-ai-chatbot

What it solves

Building a voice assistant usually requires bulky hardware and complex wiring. This project packages everything into a small device using a Raspberry Pi and a dedicated HAT, so you can carry an AI chatbot in your pocket and talk to it hands-free or with a button press.

How it works

The device uses a Raspberry Pi Zero 2w or Pi 5 with a PiSugar Whisplay HAT (which includes an LCD screen, speaker, and microphone). You press a button, speak, and the chatbot replies aloud. The project provides scripts to install dependencies, configure environment variables, build the software, and run the chatbot service. It also supports optional features like a wake word for hands-free use, image generation from text prompts, and battery level display. The setup uses a .env file for configuration, and a CLI tool (whisplay) manages settings and plugins. Pre-built images are available from the wiki for convenience.

Who it's for

This is for hobbyists and makers who want a portable, interactive AI assistant. It's suitable for anyone comfortable with Raspberry Pi and command-line setup, and it works with optional AI accelerator cards (LLM8850, Raspberry Pi AI HAT+ 2) to run models locally or via API.

Highlights

• Pocket-sized form factor with an all-in-one HAT (screen, speaker, mic)
• Simple button-press interaction with spoken replies
• Optional wake word for hands-free operation
• Image generation capability from text prompts
• Battery level display support
• Modular installation with CLI configuration wizard
• Support for offline builds on Raspberry Pi 5 with AI accelerator cards
• Pre-built OS images available for easy setup