What this site is
Six traditional games — Gomoku, Reversi, Janggi, chess, shogi and Go — each with a playable board, a tunable AI opponent, and a written guide covering the rules, the core strategic ideas, and the history of the game. There is no account system, no matchmaking queue and no chat.
How the AI works
The piece games use classical game-tree search: alpha-beta pruning over a hand-written evaluation function, with iterative deepening under a fixed time budget so the engine always answers promptly. Go uses Monte-Carlo tree search with random playouts, which is the approach that made 9×9 computer Go respectable long before neural networks arrived. All of it is TypeScript compiled to JavaScript and executed in a Web Worker on your own device.
How strong is it, honestly
It is a browser, not a data centre. Expect a strong club player in Reversi and Gomoku, a decent amateur in chess and shogi, and a mid-kyu opponent on a 9×9 Go board. That is deliberately the range where a game stays interesting for most people — you should be able to beat level 1 immediately and struggle against level 4.
Naming and trademarks
The rules of all six games are traditional and in the public domain. Where a modern trademark exists for a commercial edition of a traditional game, we use the historic generic name instead — most notably Reversi rather than the trademarked "Othello".
Contact
This is a hobby project. If something is broken, mis-translated, or a rule is implemented incorrectly, corrections are genuinely welcome.