Sno Station
Sno Station is your agents' workstation: open-source software that turns the AI agents you already run into one squad on your own machine. Two heads are better than one, and they get better with every
Free, open-source analysis for your chess games.
PluginsJavaScript · CSSGitHub @T-JulsgaardOnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Run /nomorevibe verify in your project folder to verify ownership and manage the introduction yourself.
Verify in one minuteAsk about other verification options
Free, open source chess game review. [Currently 500+ active, weekly users]
Translation pending
Taken from the repository README
Loading the full README.
Read the full README ↗Free, open-source game review for your online chess games, powered by Stockfish NNUE running locally in your browser. One click turns any Chess.com or Lichess game into a full review — accuracy scores, move-by-move classifications, an evaluation graph, and an estimated rating. Local analysis needs no account or server, and no manual PGN copying.
Learn with Chess Review opens from the button beneath Engine, beside the move controls. By default, the review panels morph into one learning workspace while the board and top bar stay in place. Settings → General → Learning experience offers Morph in place, Morph in place 2, Morph in place 3, Studio Morph, Minimal morph, Board morph, and the original Separate page layout. Each morph has its own design and timing. The transition also runs when returning to review and respects reduced-motion preferences. Morph in place 2 glides the intact board into a centered study layout and keeps Settings available alongside learning, with reserved space on desktop windows. Morph in place 3 keeps the review board size and familiar piece handling while the surrounding panels transition into learning. The learning workspace includes 23 concept lessons with 47 exercises, 25,000 offline puzzles, 1/5/10-minute challenges, and 12 opening courses with 24 annotated lines. Lessons use move, square-selection and position-reading exercises. Opening practice supports both colors and automatic opponent replies. Progress and bookmarks stay in your browser, and Back to review restores your game position and orientation. The learning objective and progress tracker preserves the full development brief for work on another computer; content sources record licensing and validation scope.
The right-hand Settings panel includes an optional Concepts tab after General and Engine. A separate Concepts panel to the right of Accuracy and Engine uses those same findings, ordered by likely relevance and phrased from your playing side. It follows the selected move, includes every match, and separates insights with space. Its top aligns with the coach commentary panel; its bottom aligns with Engine. Expand an insight's evidence to inspect the original explanation and exact scope. Accuracy and the existing move commentary retain their scoring and behavior.
The Settings Concepts panel remains the raw inspection and game-debug view. Check Enable concept analysis to list supported findings for the selected move, with exact verified scopes and game-wide debug metrics. It defaults off and stops its worker immediately when disabled. Findings use completed E080 research's bounded mechanics; broader strategic benefits and coaching usefulness remain unproven. Existing scores and coach comments are unchanged.
Source: https://github.com/T-Julsgaard/Chess-Review
Chrome webshop: https://chromewebstore.google.com/detail/chess-review/pdbffcjdmcadihmnmenkadndbdbigfam
Firefox Add-ons: https://addons.mozilla.org/en-US/firefox/addon/chess-review/
Click the image below to watch the introduction video.
Ctrl+Shift+Y.You can also paste a game URL or PGN into the popup. If your username cannot be detected, enter it once; it is remembered locally. Chess Review is intended for review after play.
Settings let you choose an engine, rating mode, board, pieces, sounds, coach and category-badge appearance. Badge numbers describe move categories; they are separate from position evaluation and game accuracy.
Both engines offer Use recorded rating and Moves only. The first compares your performance with public players near the rating saved in the game; the second ignores that rating and estimates a blitz rating level from your choices. Without a saved rating, the review falls back to Moves only, which needs at least ten nonforced decisions. Each engine uses its own models.
Choose Settings → General → Interface presets → Review experience → Dopamine for a more expressive review: purple board and interface, move-colored accents, and reactions centered on the move's destination. Brilliant moves send teal waves across the squares; blunders produce a red shockwave and shards. Other categories have smaller reactions. Effects never cover the board with a text box or intercept input. The same settings section offers Preview Brilliant effect and Preview Blunder effect buttons, without changing your move's actual classification.
The commentary explains available engine evidence: immediate mate, a mating evaluation, a legal capture, material changes in the shown line, or the evaluation loss when no specific tactic is established. You can explore the reply and the engine's better alternative. Explanations use existing local analysis; a shown continuation is not a promise that every future move is forced. Missing or invalid lines never produce invented tactical claims.
Normal remains the default. The mode and original visual preferences are saved separately; switching back restores the original appearance and commentary. The default Modern layout gives the explanation panel more room only in Dopamine mode; custom arrangements stay intact. Reduced motion or disabling Animation keeps a static destination outline. No extra engine searches, network services, extension permissions, dependencies, or changes to classifications and accuracy are needed.
For just the explanation text with your usual appearance, turn on Settings → General → Coach → Engine explanations. It works with any coach, including No coach, and saves your choice independently of the review mode. Engine explanations and Special replies are alternative reply styles: turning either on switches the other off. Dopamine keeps engine explanations by default until you explicitly choose a reply style in Coach settings.
Chess Review is open source and uses local Stockfish analysis with documented scoring models. The repository includes the fitting tools, public-data evidence and reference results needed to inspect and reproduce the bundled numerical models.
With Node.js 24 or later, reproduce the bundled coefficients and reference scores offline:
node tools/calibration/reproduce-public.mjs
This check replays the included evidence without rerunning engine searches. The methodology documents the inputs, calculations and limitations. Estimated ratings are approximate.
Move categories follow documented classification rules. See the special annotations guide for their definitions and sources, and the repository guide for the supporting files.
The separate research workspace investigates improvements to accuracy, estimated ratings and move categories against our own documented baseline. It records reproducible experiments, successful and unsuccessful findings, evidence limits and implementation decisions. Research candidates become part of the extension only through a documented promotion. See the research index for current state and next questions.
Install from the Chrome Web Store or Firefox Add-ons using the links above. For a local Chrome installation:
chrome://extensions and enable Developer mode (top-right).Games are fetched from Chess.com's and Lichess's public APIs and analyzed locally with bundled WebAssembly builds of Stockfish. Lookup requests include public player usernames or game IDs. No games or analysis are sent to the developer or analytics services. Firefox discloses the browsing activity and website content used for these requests. Sharing a game copies an encoded, unencrypted PGN/metadata link to your clipboard; anyone you send it to can read that information. See PRIVACY.md.
Optional puzzle leaderboards have a Cloudflare Workers + D1 service, server-verified ranked play and matching screens for all three modes. They are disabled until a deployment is configured. See leaderboard setup and security. Participation requires explicit opt-in before viewing or uploading. Nationality is optional; Settings → General → Leaderboard provides sharing, opt-out and online deletion controls. Ordinary offline progress stays local.
The Maintained source workflow runs npm run verify:source and npm run verify:history
before the test suites. The history check inspects reachable commits and historical text
blobs in batches targeting 16 MiB, so growing research history does not overflow one
combined subprocess output buffer. All existing content and path checks remain active.
The complete research suite runs in a separate Windows CI job with the original Python runtime bound by the E211 saved observations. Setup verifies the pinned archive, retained executable hash, and full build version; existing research evidence and checks remain unchanged. See runtime setup.
Run npm ci, npm test, npm run verify:engines, and npm run build.
The build creates separate Chrome and Firefox ZIPs under web-ext-artifacts/,
with browser-specific manifests and no development dependencies or tests. Each build
has its own release directory, source snapshot, and SHA-256 package checksums.
See RELEASE.md for store reviewer notes and validation commands.
Chess Review's own code is licensed under the GNU General Public License v3.0 (see
LICENSE), matching the bundled Stockfish engine. Bundled third-party assets (engine,
pieces, sounds, opening book, libraries) keep their own licenses — see
ATTRIBUTIONS.md and THIRD_PARTY_NOTICES.md.
Some asset redistribution rights remain unresolved as documented in the attributions;
the project's GPL license does not grant rights to those assets.
Corresponding source for the GPL/AGPL components is available
from the upstream projects listed there.
Disclaimer: Chess Review is an independent, unofficial tool. It is not affiliated with, endorsed by, or sponsored by Chess.com or Lichess. Those names are used only to describe the sites it reads games from (nominative reference); all trademarks belong to their respective owners.