A wearable AI system that helps new poker players learn the game in real time using a Raspberry Pi, camera-based object detection, and an LLM backend, with strategy guidance and explanations delivered through a mobile app.
Overview
The system runs on a Raspberry Pi worn by the player. A camera captures the table view and feeds it into an object-detection pipeline to identify cards, positions, and game state. That information is sent to an LLM backend, which generates real-time strategy advice, hand explanations, and teaching tips. The player receives the guidance through a companion mobile app, so they can learn while playing without leaving the table.
Tech stack
Raspberry Pi for the wearable unit, camera-based object detection for card and table recognition, an LLM service for strategy and explanation generation, and a mobile app for audio and visual feedback. The pipeline is designed for low latency so feedback stays relevant during live play.
Demo
Gallery
