ChessAgine MCP

What it does

ChessAgine MCP turns any MCP-capable AI assistant into a deep chess analyst. It renders interactive chessboard positions and full PGN game replays directly in the chat, runs Stockfish evaluation, queries Lichess for player games, and integrates neural chess engines including Maia2, Leela Chess Zero, and Elite Leela. Agents can reason about positions, compare engine evaluations, explore openings, and analyse entire games with visual board state.

Key features

• Board rendering — renders real-time chessboard positions as interactive UI resources inside the client
• PGN viewer — visualise complete games move-by-move with in-client replay
• Multi-engine analysis — compare Stockfish, Maia2, Leela, and Elite Leela evaluations on the same position
• Lichess integration — pull any player's recent games and analyse them on the spot
• Opening database — explore opening repertoires via the Chessboard Magic opening library
• Chess theme support — customisable board themes

Installation

Claude Desktop (recommended — MCPB file):

1. Download chessagine-mcp.mcpb from GitHub Releases
2. Open Claude Desktop → Settings → Extensions → Install from file
3. Select the .mcpb file and restart

Claude Desktop (local build):

git clone https://github.com/jalpp/chessagine-mcp.git
cd chessagine-mcp && npm install && npm run build

Then add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "chessagine-mcp": {
      "command": "node",
      "args": ["/absolute/path/to/chessagine-mcp/build/runner/stdio.js"]
    }
  }
}

Vercel (self-hosted HTTP): Fork the repo and deploy to Vercel — no env vars needed. Server available at https://your-project.vercel.app/mcp.

Supported hosts

Claude Desktop (primary, confirmed in README with detailed setup guide).