MCP server for playing chess against LLMs – exposes tools such as get_board_visualization, make_move, new_game etc.
https://github.com/jiayao/mcp-chessStop copy-pasting PGN files or describing chess positions in text. This MCP server gives your AI direct access to a full chess engine with visual board representation, letting you play, analyze, and test chess scenarios in real-time.
Ever tried explaining a chess position to an AI? "The knight is on f3, the bishop threatens the queen..." It's painful. With MCP Chess, your AI can see the actual board, make legal moves, and understand position dynamics instantly.
This isn't just about playing games - it's about having a proper chess environment for:
Visual Board State: Every position rendered as an image that automatically flips based on your color. No more "imagine the board from Black's perspective."
Full Game Management: Start games, make moves in standard notation, get legal move validation, track game status. Everything you need for complete chess sessions.
PGN Analysis: Drop in a PGN string and find specific positions based on piece placement. Perfect for analyzing famous games or debugging chess algorithms.
Smart Move Handling: Works with both algebraic notation (Nf3, Bxe5) and UCI format (g1f3, c1e5). The AI doesn't need to worry about notation conversion.
AI Chess Training: Test different models' chess understanding by having them play against each other or analyze positions. The visual feedback makes it immediately clear when an AI misunderstands the position.
Educational Content: Build chess tutorials or explanations where the AI can show positions, demonstrate tactics, and walk through game analysis with actual board images.
Game Development: Prototype chess features, test game logic, or validate chess algorithms against a proven engine without building your own chess infrastructure.
Research & Analysis: Study opening theory, endgame positions, or tactical patterns with an AI that can see and manipulate the actual board state.
Add to your MCP config:
{
"mcpServers": {
"chess": {
"command": "uvx",
"args": ["mcp-chess"]
}
}
}
That's it. Your AI now has access to new_game(), make_move(), get_board_visualization(), and position analysis tools.
Instead of building chess logic from scratch or dealing with multiple chess libraries, you get a complete chess environment that speaks directly to your AI. The visual board representation means you can actually see what's happening, making debugging and testing infinitely easier.
Whether you're building chess applications, testing AI capabilities, or just want to play a proper game with visual context, this server handles the complexity while keeping the interface clean and intuitive.
Ready to give your AI actual chess vision? Install it in 30 seconds and start playing.