PostMatem

Client-Side Chess Workstation and Game Review Studio

---

Overview

PostMatem is a private, client-side chess workstation providing real-time game reviews, blunder workouts, branching move tree analysis, and engine sparring. The platform runs entirely within the browser via WebAssembly (WASM) and multi-threaded Web Workers, requiring no server-side computation, user accounts, or telemetry.

Features

• Client-Side Engine Execution: Stockfish 19 and 17 Lite NNUE run locally inside sandboxed Web Workers with no external evaluation requests.
• Game Review and Move Classification: Move classification based on an Expected Points Lost ($\Delta\text{XP}$) model, detecting Brilliant sacrifices, Great moves, Best moves, Inaccuracies, Mistakes, Blunders, and Misses.
• Blunder Workout: Interactive puzzle generator that rewinds the board to critical errors, tests refutations against Stockfish, and provides progressive hint tiers.
• Advantage and Win-Rate Chart: Continuous win-percentage curves over the course of the game with interactive move scrubbing.
• Versus Engine Sparring: 20 calibrated Elo rating tiers across Casual (hints, evaluation bar, takebacks) and Competitive (strict touch-move rules) modes.
• ECO Opening Book: Searchable catalog of over 3,000 opening variations with one-click analysis board loading.
• External Match Importer: Direct PGN retrieval from Lichess and Chess.com public user feeds.
• Export Formats: Standard annotated PGN export with Informant NAG glyphs, and 1200x630px social summary cards.
• Local Persistence: Offline game archiving and evaluation caching powered by IndexedDB.

Keyboard Shortcuts

Key	Action
Right / Space	Advance to next move
Left	Return to previous move
Up / Home	Jump to start of game
Down / End	Jump to end of game
Z / Ctrl + Z	Takeback move (Casual mode)
F	Flip board perspective
E	Toggle evaluation bar
H	Toggle best-move suggestion arrow
B	Toggle move classification badges

Architecture & Tech Stack

• Framework: React 19, TypeScript, Vite
• State Management: Zustand and Immer
• Chess Logic: chess.js
• Board Component: react-chessboard
• Engine: Stockfish 19 Lite via WebAssembly
• Storage: idb (IndexedDB wrapper)
• Iconography: Lucide React

Getting Started

Prerequisites

• Node.js 18 or higher (Node.js 20 LTS recommended)
• npm, pnpm, or yarn

Installation

git clone https://github.com/M-Nikox/PostMatem.git
cd PostMatem
npm install

Development Server

npm run dev

The application will be accessible at http://localhost:5173.

Production Build

npm run build
npm run preview

Optional Full NNUE Engine (Local Setup)

PostMatem defaults to the single-threaded Stockfish 19 Lite WASM binary (~1.7 MB). To copy and enable the uncompressed 99 MB full neural network binary locally:

npm run setup:engines:full

Deployment

PostMatem produces static assets compatible with any static hosting provider. The build output requires standard WebAssembly cross-origin isolation headers (Cross-Origin-Opener-Policy: same-origin and Cross-Origin-Embedder-Policy: credentialless).

• GitHub Pages: Automated deployment is configured in .github/workflows/deploy.yml.
• Cloudflare Pages: Set build command to npm run build and output directory to dist. Preconfigured via public/_headers.
• Vercel: Configuration and security headers are specified in vercel.json.
• Netlify: Configuration and security headers are specified in netlify.toml.

Credits and Attributions

PostMatem incorporates work from the open-source chess and software community:

• Stockfish: Chess engine by Marco Costalba, Joona Kiiski, Tord Romstad, Stéphane Nicolet, and the Stockfish development community (GPLv3).
• Stockfish.js: WebAssembly engine port by nmrugg and contributors (GPLv3).
• CentiChess: Architecture reference for Web Worker engine pools by Cooper Ross (CC0 1.0).
• Chesskit: Reference architecture by GuillaumeSD.
• Lichess: Mathematical win-rate model and ECO dataset (AGPLv3 / CC0).
• Colin M.L. Burnett: Staunton vector chess pieces (CC BY-SA 3.0 / GFDL).
• RhosGFX: Themed piece artwork and board illustrations.

Detailed license information and original links are available in CREDITS.md.

Author

• M-Nikox — @M-Nikox

License

PostMatem is free software licensed under the GNU General Public License v3.0 (GPLv3). Stockfish engine binaries in public/engines/ are distributed under the GNU General Public License v3.0 (GPLv3).