Real-time AI Chess Analysis

This project combines computer vision, machine learning, and a self-hosted chess engine to provide real-time analysis of over-the-board chess games.

Using a camera positioned above the chessboard, the system detects every move played during a live game. After each move, the current board position is automatically reconstructed and sent to a self-hosted chess engine, which immediately evaluates the position and returns an analysis, allowing players to follow the game with professional-level insights in real time.

Once the game is finished, the application automatically generates and copies the complete PGN notation, making it easy to build a personal archive of games and review previous matches.

I also created a detailed video explaining the entire development process, including the computer vision pipeline, machine learning model, and the overall system architecture. You can watch it on YouTube using the link below.

▶

Watch the full project breakdown on YouTube.