Show HN: Maia Chess – Human-like chess AI for playing, learning, and more

Play Against Maia

Play against the most human-like chess AI

Challenge Maia, a neural network trained to play like a human at your chosen rating level. Unlike traditional engines that play robotically, Maia naturally plays moves that a person would make.

Trained on millions of human games, Maia plays with human chess intuition and decision-making style.

Game Analysis

Analyze games with human-aware AI

Go beyond perfect engine lines—see what real players would actually do. Maia combines Stockfish's precision with human tendencies learned from millions of real games, giving you real-world context in every position. Instantly tell whether a move wins only for computers or also works at your rating, and where players like you are most likely to stumble.

Explore the top moves at every rating level, spot positions where blunders are likely, and understand how to level up your play in every single position. Get personalized insights based on your playing style and rating level.

Human-Centered Puzzles

Train with Maia as your coach

Maia curates puzzles based on its understanding of how millions of players improve. With Maia puzzles, you can benchmark your vision, focus on your gaps in understanding, and turn hard-to-spot ideas into second nature.

Each puzzle includes data showing how players of different ratings approach the position, making your training more targeted and effective.

Maia is a human-like chess engine, designed to play like a human instead of playing the strongest moves. Maia uses the same deep learning techniques that power superhuman chess engines, but with a novel approach: Maia is trained to play like a human rather than to win.

Maia is trained to predict human moves rather than to find the optimal move in a position. As a result, Maia exhibits common human biases and makes many of the same mistakes that humans make. We have trained a set of nine neural network engines, each targeting a specific rating level on the Lichess.org rating scale, from 1100 to 1900.

We introduced Maia in our paper that appeared at KDD 2020, and Maia 2 in our paper that appeared at NeurIPS 2024.