Ask why a move was a mistake, or what you should have played instead, and get a straight answer in plain words. Snowie, your AI coach, is built on a full Stockfish review of your game, so every explanation is grounded in the real lines the engine saw rather than guessed.
Runs locally on your machine, and the AI coach uses your existing Claude subscription. No API keys, no per-token bills.
A raw engine spits out a number and a best move. That's great if you already understand the position, and useless if you don't. Snowie, the built-in AI coach, turns that number into a sentence.
Best move: c3. …and you're left to figure out the rest yourself.
“Nf3 was a blunder. Your win chance fell from 81% to 57%. The crushing reply was c3!, kicking the knight with nowhere safe to go. Instead, Nf3 invited …Nxf3+ and handed back the whole advantage.”
Everything is built from one Stockfish sweep of your game: no guessing, no estimates.
Every inaccuracy, mistake, and blunder gets a concrete comment: the better move, its line, and how your move gets punished.
Replay each mistake, drag a piece to try your own idea, and free-explore any variation with instant engine feedback.
A Lichess-style win graph across the whole game, oriented to the side you played. Click or arrow-key to scrub.
See the move you played, the engine's best moves (live multi-PV), and the refutation of any move you try.
Ask Snowie “why is this bad?” or “what should I do here?” and it answers from pre-computed engine facts, so it stays grounded.
Every game is saved locally and tagged for recurring weaknesses (hung pieces, missed forks, time trouble) that build into your profile.
Pull games straight from Lichess, paste a PGN from anywhere, or bulk-upload a Chess.com export, then walk through them move by move.
One process holds one Stockfish engine and one analysis. A Claude Code terminal (via MCP) and an interactive browser board both read from it, so they never disagree.
Works with games from anywhere — Lichess, Chess.com, or any PGN. It runs locally and sets itself up: no Python, no Homebrew, nothing to install first — it fetches Stockfish and the rest on first launch.
Download Tintin's AI Chess Analysis.app (from Releases), drag it into Applications, and open it (first time, unsigned: double-click, then System Settings → Privacy & Security → Open Anyway). It installs everything itself — no Homebrew or Python needed — then opens the board; paste any PGN, or enter a Lichess username to load your recent games.
Double-click Tintin's AI Chess Analysis.bat (Windows) or Tintin's AI Chess Analysis.command (Linux). The first launch installs everything, then opens the board — paste any PGN, or load your recent Lichess games.
Run ./install.sh (or install.ps1 on Windows) and reload Claude
Code. Paste a PGN, say “analyze this game,” and Claude narrates your mistakes and hands
you the board link.
Full setup, configuration, and options are in the README on GitHub.
claude CLI, and how is it different from the Claude app?
The claude CLI (Claude Code) is a separate program that runs in your computer's
terminal. It is not the Claude desktop app or the claude.ai website, but it
signs in with the same Claude subscription, so there is no extra account, no
API key, and no per-token bill. This app uses the CLI behind the scenes to write the
plain-English coaching, which is why you need it installed and logged in.
Open a terminal and install Claude Code with one of these (the native install is recommended):
# macOS / Linux / WSL
curl -fsSL https://claude.ai/install.sh | bash
# Windows (PowerShell)
irm https://claude.ai/install.ps1 | iex
Or use a package manager: brew install --cask claude-code (macOS) or
winget install Anthropic.ClaudeCode (Windows).
Then run claude once and follow the prompt to log in with your Claude
subscription. After that, restart this app and the AI coach will work. Full details are in
the Claude Code docs.
Note: the engine review itself (the mistake list, eval bar, win graph, and arrows) works without the CLI. You only need it for the conversational coaching and the AI summaries.