당신이 실제로 할 수있는 체스 코치 ** 톡 **, 그리고 그 ** 당신은 일을 만들 수 없습니다 **. 왜 움직이는가? 실수 또는 당신이 연주해야 할 일, 그리고 스트레이트 대답 **에서 단어 **, 접지 real **Stockfish ** 추측 대신 라인. 그 후, 그것은 엔진과 게임을 검토하고 잘못되었는지 정확히 찾을 수 있습니다. 차이는 그런 다음 설명합니다. 에서 게임 ** (Lichess, Chess.com, 또는 당신이 풀 수 있는 어떤 PGN). Lichess와 Chess.com 게임 모두 할 수 있습니다 으로 fetched by 사용자 이름, 최신 게임 자동 로드 출시, 그리고 새로운 **Chess.com 게임 동기화 자동으로 ** 당신의 역사에. Lichess 선수는 몇 가지 여분을 얻을. 그것은 두 가지 방법을 실행: Claude 코드에서 터미널(MCP 서버) 및 interactive web board로 한 엔진과 하나의 공유 분석, 그래서 그들은 결코 동의하지 않습니다.

그것에는 ** 2개의 형태 **, 동일한 널에 의하여 측이 있습니다:

• Analyze 당신의 게임: 위에 설명된 전체 게임 검토.
• ** 퍼즐 ** : 충실한 ** Glicko-2** 등급을 가진 내장 전술 트레이너 ** rush, and, uniquely, AI 코치 각 솔루션 뒤에 패턴를 태워. 그것은 또한 할 수 있습니다 blunders**를 연습 퍼즐으로 바꾸십시오.

Chess Review MCP

A chess coach you can actually talk to, and one that doesn’t make things up. Ask why a move was a mistake or what you should have played, and get a straight answer in words, grounded in real Stockfish lines instead of guessed. Under the hood it reviews your game with the engine and finds exactly where you went wrong; the difference is it then explains it. Works with games from anywhere (Lichess, Chess.com, or any PGN you can paste). Both Lichess and Chess.com games can be fetched by username, your latest game auto-loads on launch, and new Chess.com games sync automatically into your history. Lichess players get a few extras. It runs two ways: from the Claude Code terminal (as an MCP server) and as an interactive web board that share one engine and one analysis, so they never disagree.

It has two modes, side by side in the same board:

• Analyze your games: the full-game review described above.
• Puzzles: a built-in tactics trainer with a faithful Glicko-2 rating, a timed Storm rush, and, uniquely, an AI coach that teaches the pattern behind each solution. It can also turn your own past blunders into practice puzzles. See Puzzle trainer.

Requires a Claude subscription (or a local model). The AI-coach explanations and chat (Snowie), the whole “explained in words” part, run on headless claude, using your existing Claude subscription (no API key, no per-token billing). You need the claude CLI installed and logged in (claude login). Prefer your own model? You can instead point the AI coach at a local model (Ollama, LM Studio, …) or any OpenAI-compatible provider (with its API key), and use no Claude account at all; see Run the AI on your own model. Either way, the engine review itself (mistake list, eval bar, win graph, arrows) works without any of it.

New here? Pick your goal:

• 🎯 I just want to review my games (and train puzzles) → get the app (download on Mac, or double-click the launcher on Windows/Linux; it sets itself up).
• 🤖 I want chess analysis inside Claude Code → run the installer.