erlangzhang/alan-aumaton

alan-aumaton — move API

Serves moves from a chess policy network trained to imitate a specific player. Backend for an interactive board hosted elsewhere (e.g. GitHub Pages).

• —GET /health → {status, username, model_loaded}
• —POST /move {fen, temperature} → {move, san, fen, is_game_over, result, top_moves}

Built with FastAPI + PyTorch (CPU). See the project repo for training code.