Stockfish
Enables AI to communicate with the Stockfish chess engine.
mcp-stockfish 🐟
A Model Context Protocol server that lets your AI talk to Stockfish. Because apparently we needed to make chess engines even more accessible to our silicon overlords.

🧠⚡🖥️ Your LLM thinks, Stockfish calculates, you pretend you understand the resulting 15-move tactical sequence.
What is this?
This creates a bridge between AI systems and the Stockfish chess engine via the MCP protocol. It handles multiple concurrent sessions because your AI probably wants to analyze seventeen positions simultaneously while you're still figuring out why your knight is hanging.
Built on mark3labs/mcp-go. Because reinventing wheels is for people with too much time.
Features
- 🔄 Concurrent Sessions: Run multiple Stockfish instances without your CPU crying
- ⚡ Full UCI Support: All the commands you need, none of the ones you don't
- 🎯 Actually Works: Unlike your last side project, this one has proper error handling
- 📊 JSON Everything: Because apparently we can't just use plain text anymore
- 🐳 Docker Ready: Containerized for when you inevitably break your local setup
Supported UCI Commands ♟️
| Command | Description |
|---|---|
uci | Initializes the engine in UCI mode |
isready | Checks if the engine is ready. Returns readyok |
position startpos | Sets up the board to the starting position |
position fen [FEN] | Sets up a position using FEN notation |
go | Starts the engine to compute the best move |
go depth [n] | Searches n plies deep. Example: go depth 10 |
go movetime [ms] | Thinks for a fixed amount of time in milliseconds. Example: go movetime 1000 |
stop | Stops current search |
quit | Closes the session |
Quick Start
Installation
git clone https://github.com/sonirico/mcp-stockfish
cd mcp-stockfish
make install
Usage
# Default mode (stdio, because we're old school)
mcp-stockfish
# With custom Stockfish path (for the special snowflakes)
MCP_STOCKFISH_PATH=/your/special/stockfish mcp-stockfish
# HTTP mode (for the web-scale crowd)
MCP_STOCKFISH_SERVER_MODE=http mcp-stockfish
Configuration ⚙️
Environment Variables
Server Configuration
MCP_STOCKFISH_SERVER_MODE: "stdio" or "http" (default: "stdio")MCP_STOCKFISH_HTTP_HOST: HTTP host (default: "localhost")MCP_STOCKFISH_HTTP_PORT: HTTP port (default: 8080)
Stockfish 🐟 Configuration
MCP_STOCKFISH_PATH: Path to Stockfish binary (default: "stockfish")MCP_STOCKFISH_MAX_SESSIONS: Max concurrent sessions (default: 10)MCP_STOCKFISH_SESSION_TIMEOUT: Session timeout (default: "30m")MCP_STOCKFISH_COMMAND_TIMEOUT: Command timeout (default: "30s")
Logging
MCP_STOCKFISH_LOG_LEVEL: debug, info, warn, error, fatalMCP_STOCKFISH_LOG_FORMAT: json, consoleMCP_STOCKFISH_LOG_OUTPUT: stdout, stderr
Tool Parameters
command: UCI command to executesession_id: Session ID (optional, we'll make one up if you don't)
Response Format
{
"status": "success|error",
"session_id": "some-uuid",
"command": "what you asked for",
"response": ["what stockfish said"],
"error": "what went wrong (if anything)"
}
Session Management
Sessions do what you'd expect:
- Spawn Stockfish processes on demand
- Keep UCI state between commands
- Clean up when you're done (or when they timeout)
- Enforce limits so you don't fork-bomb yourself
Integration
Claude Desktop
{
"mcpServers": {
"chess": {
"command": "mcp-stockfish",
"env": {
"MCP_STOCKFISH_LOG_LEVEL": "info"
}
}
}
}
Development
make deps # Get dependencies
make build # Build the thing
make test # Run tests (when they exist)
make fmt # Make it pretty
Credits 🐟
Powered by Stockfish, the chess engine that's stronger than both of us combined. Created by people who actually understand chess, unlike this wrapper.
Thanks to:
- The Stockfish team for making chess engines that don't suck
- MCP SDK for Go for handling the protocol so I don't have to
- Coffee
License
MIT - Do whatever you want, just don't blame me when it breaks.
Related Servers
Neume
Make songs with AI
root-mcp
MCP server for ROOT CERN files
SciPilot
Natural language interface for scientific command-line tools via MCP
WhisperGraph MCP
Open-source, self-hostable MCP server for WhisperGraph — a graph of 7.39B nodes / 39B edges mapping DNS, BGP, GeoIP, WHOIS, and threat intelligence. Six read-only tools (Cypher query + schema introspection + threat assessment), six resources, eight investigation prompts. stdio and Streamable HTTP transports.
mcp-fns-check
MCP-сервер для проверки российских контрагентов (юридические лица и ИП) через публичные данные ФНС: ЕГРЮЛ/ЕГРИП, ЕФРСБ, «Прозрачный бизнес», ФССП, КАД.
Zadig&Voltaire
Place order on Zadig&Voltaire website
OctoEverywhere For 3D Printing
A 3D Printing MCP server that allows for querying for live state, webcam snapshots, and 3D printer control.
Public Data Portal Short-term Forecast
Provides current weather information using the Korea Meteorological Administration's short-term forecast API.
TradeMemory Protocol
AI trading memory layer for MT5/forex with 15 MCP tools — store/recall trades, pattern discovery, strategy evolution, and Outcome-Weighted Memory.
Memory Bank MCP
A production-ready Model Context Protocol (MCP) server that provides a powerful, vector-native memory bank for AI agents. Built with the Protocol-Lattice Go Agent Framework, this server offers persistent, searchable, and shareable memory with multiple database backends.