First complete Rust version of Kłak, a UCI chess engine built as a learning project.
Following the controversy surrounding my previous engine, Luna CE, specifically regarding the use of AI in its development, I decided to learn Rust and write an engine entirely on my own.
It has Negamax with alpha-beta, quiescence search, MVV-LVA move ordering, piece-square tables, iterative deepening with time management.
Known limits: no repetition detection, increment ignored, no position fen, one-move PV.
https://github.com/Spunc595/Klak/releases/tag/v1.0.0
Introducing Kłak v1.0.0
Moderator: Ras
-
Spunc595
- Posts: 48
- Joined: Mon Jul 06, 2026 12:15 am
- Full name: Daniele Marpino
-
Spunc595
- Posts: 48
- Joined: Mon Jul 06, 2026 12:15 am
- Full name: Daniele Marpino
Re: Introducing Kłak v1.0.0
This is my son's engine.
GrapeChessEngine
A small UCI chess engine written in Python, built step by step on top of python-chess. Made by Alessandro Marpino. License: MIT. The Windows build includes PyPy and python-chess (see their licenses in the licenses folder).
Search: negamax with alpha-beta, quiescence search, iterative deepening with time management, transposition table, MVV-LVA move ordering (TT move first), cheap draw detection.
Evaluation: material + piece-square tables (precomputed, read from bitboards), tapered king evaluation (smooth blend between middlegame and endgame tables based on the material left).
Speed: around 100k nodes/s with PyPy.
Protocol: UCI, works in Arena, Cute Chess and other GUIs.
Play it on Lichess: GrapeChessEngine
https://github.com/alemarpino-rgb/Grape ... g/Grape080
https://lichess.org/@/GrapeChessEngine
GrapeChessEngine
A small UCI chess engine written in Python, built step by step on top of python-chess. Made by Alessandro Marpino. License: MIT. The Windows build includes PyPy and python-chess (see their licenses in the licenses folder).
Search: negamax with alpha-beta, quiescence search, iterative deepening with time management, transposition table, MVV-LVA move ordering (TT move first), cheap draw detection.
Evaluation: material + piece-square tables (precomputed, read from bitboards), tapered king evaluation (smooth blend between middlegame and endgame tables based on the material left).
Speed: around 100k nodes/s with PyPy.
Protocol: UCI, works in Arena, Cute Chess and other GUIs.
Play it on Lichess: GrapeChessEngine
https://github.com/alemarpino-rgb/Grape ... g/Grape080
https://lichess.org/@/GrapeChessEngine