Babelfish

Babelfish 🐟

"Speak to your chess engine"

Babelfish is a chess analysis tool that uses Stockfish via MCP (Model Context Protocol) to provide context-rich analysis of chess positions and games. It bridges the gap between powerful chess engines and language models, making chess analysis more accessible and conversational.

Features

• 📊 Position Analysis - Deep analysis of chess positions using Stockfish
• 🎯 Move Suggestions - Get the best moves with explanations
• 📚 Game Analysis - Analyze complete games move by move
• 🗣️ Human-Readable Explanations - Convert engine output into natural language
• 🔌 MCP Integration - Works seamlessly with Claude Desktop and other MCP clients

Installation

This project uses uv for dependency management:

# Clone the repository
git clone <repository-url>
cd babelfish

# Install dependencies
uv sync

# Install Stockfish engine
# Ubuntu/Debian:
sudo apt install stockfish

# macOS:
brew install stockfish

# Windows:
# Download from https://stockfishchess.org/download/

Usage

Web Interface (Recommended)

The most user-friendly way to use Babelfish is through its web interface. This provides an intuitive chat-based interface for chess analysis.

Quick Start

1. Set up your OpenRouter API key:

   export OPENROUTER_API_KEY="your-api-key-here"
   

2. Start the web interface:

   uv run python start_web.py
   

3. Open your browser and go to http://localhost:5000

Model Recommendations and Costs

⚠️ Important Cost Warning: Using OpenRouter models incurs costs based on token usage. Monitor your usage to avoid unexpected charges.

Recommended Model: Based on testing, openai/o1 provides the best chess analysis quality, offering superior positional understanding and tactical insights compared to other models.

Alternative Models: You can also try:

• anthropic/claude-3.5-sonnet - Good balance of cost and quality
• openai/gpt-4o - Strong general performance
• meta-llama/llama-3.1-70b-instruct - More cost-effective option

Features

• Interactive Chat Interface: Ask questions about positions in natural language
• Visual Chess Board: See positions and moves clearly displayed
• Real-time Analysis: Get instant feedback on positions and moves
• Game Analysis: Upload PGN files or input move sequences
• Multiple Model Support: Choose from various AI models via OpenRouter

Example Usage

Once the web interface is running, you can:

• Analyze a position: "What's the best move in this position: rnbqkbnr/pppp1ppp/8/4p3/4P3/8/PPPP1PPP/RNBQKBNR w KQkq e6 0 2"
• Evaluate moves: "Is Nf3 a good move here?"
• Get strategic advice: "What's the plan for White in this position?"
• Analyze games: Paste PGN notation or describe a game sequence

Standalone Demo

Run the demo to test the chess analyzer:

python main.py

MCP Server Mode

Start the MCP server for integration with other tools:

python main.py --mcp

Claude Desktop Integration

Standard Setup (Same System)

To use Babelfish with Claude Desktop in stdio mode on the same system:

1. Install Stockfish (see installation section above)

2. Configure Claude Desktop by adding this to your claude_desktop_config.json:

{
  "mcpServers": {
    "babelfish": {
      "command": "uv",
      "args": ["run", "python", "main.py", "--mcp"],
      "cwd": "/path/to/your/babelfish"
    }
  }
}

Note: Replace /path/to/your/babelfish with the actual path to your babelfish directory.

WSL + Windows Setup

To run Babelfish in WSL while using Claude Desktop on Windows:

1. Install Stockfish in WSL (see installation section above)

2. Configure Claude Desktop on Windows by adding this to your claude_desktop_config.json:

{
  "mcpServers": {
    "babelfish": {
      "command": "wsl",
      "args": ["-d", "Ubuntu", "bash", "-c", "cd ~/babelfish && uv run python main.py --mcp"],
      "cwd": "C:\\"
    }
  }
}

Important Notes for WSL Setup:

• Replace Ubuntu with your WSL distribution name if different (check with wsl -l in Windows PowerShell)
• Replace ~/babelfish with the actual path to your babelfish directory in WSL
• Make sure WSL is set up and your distribution is running
• Ensure uv and all dependencies are installed in the WSL environment, not Windows

Alternative WSL Configuration (if you have a specific WSL distribution):

{
  "mcpServers": {
    "babelfish": {
      "command": "wsl.exe",
      "args": ["--distribution", "Ubuntu-22.04", "--exec", "bash", "-c", "cd ~/babelfish && uv run python main.py --mcp"]
    }
  }
}

3. Configuration File Locations:

   • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
   • Windows: %APPDATA%\Claude\claude_desktop_config.json
   • Linux: ~/.config/Claude/claude_desktop_config.json

4. Restart Claude Desktop after adding the configuration.

5. Verify Installation - You should see the Babelfish tools available in Claude Desktop:

   • analyze_position - Analyze a chess position
   • analyze_game - Analyze a complete game
   • explain_position - Get human-readable position explanations

Example Usage in Claude Desktop

Once configured, you can ask Claude to analyze chess positions:

"Can you analyze this chess position: rnbqkbnr/pppp1ppp/8/4p3/4P3/8/PPPP1PPP/RNBQKBNR w KQkq e6 0 2"

Or analyze complete games:

"Please analyze this game: 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6"

Available MCP Tools

analyze_position

Analyzes a chess position using the Stockfish engine.

Parameters:

• fen (required): Chess position in FEN notation
• depth (optional): Analysis depth (default: 15)

analyze_game

Analyzes a complete chess game move by move.

Parameters:

• moves (required): Array of moves in standard algebraic notation

explain_position

Provides human-readable explanations of chess positions.

Parameters:

• fen (required): Chess position in FEN notation

Requirements

• Python 3.8+
• Stockfish chess engine installed on your system
• uv package manager

Troubleshooting

"Stockfish not found" error

Make sure Stockfish is installed and accessible from your PATH:

# Test if Stockfish is available
stockfish

If not available, install using your system's package manager or download from stockfishchess.org.

MCP Server not connecting

1. Verify the path in your claude_desktop_config.json is correct
2. Make sure you can run uv run python main.py --mcp from the command line
3. Check Claude Desktop logs for error messages
4. Restart Claude Desktop after configuration changes

WSL-specific troubleshooting

WSL Distribution Issues

# Check available WSL distributions
wsl -l -v

# Make sure your distribution is running
wsl -d Ubuntu

# Test the command from Windows PowerShell
wsl -d Ubuntu bash -c "cd ~/babelfish && uv run python main.py --mcp"

Path and Permission Issues

# In WSL, verify the project path exists
ls -la ~/babelfish/

# Make sure uv is installed in WSL
which uv

# Test stockfish in WSL
which stockfish
stockfish

Common WSL Setup Issues

• WSL not installed: Install WSL2 from Microsoft Store or PowerShell: wsl --install
• Distribution not running: Start it with wsl -d Ubuntu (replace Ubuntu with your distro)
• uv not found in WSL: Install uv in WSL environment, not Windows
• Permission errors: Ensure the WSL user has access to the babelfish directory
• Stockfish not available: Install stockfish in WSL: sudo apt install stockfish

Testing the WSL Integration

From Windows Command Prompt or PowerShell:

# Test if WSL can access the project
wsl ls ~/babelfish/

# Test if the command works
wsl -d Ubuntu bash -c "cd ~/babelfish && python main.py"

Contributing

Contributions are welcome! Please feel free to submit issues and pull requests.

License

This project is open source. Please check the LICENSE file for details.