Xplainable MCP Server

# Xplainable MCP Server - Claude Desktop Setup This guide shows you how to add the Xplainable MCP Server to Claude Desktop using a Python virtual environment for reliable dependency management. ## Setup Instructions ### 1. Get Your API Key Get your Xplainable API key from: https://platform.xplainable.io ### 2. Create Python Virtual Environment **Create and set up the virtual environment:** ```bash # Navigate to your projects directory cd /path/to/your/projects # Clone the MCP server (if you haven't already) git clone https://github.com/xplainable/xplainable-mcp-server.git cd xplainable-mcp-server # Create Python virtual environment python3 -m venv xplainable-mcp-env # Activate the environment source xplainable-mcp-env/bin/activate # Install the MCP server and dependencies pip install -e . # Verify installation python -m xplainable_mcp.server --help ``` ### 3. Configure Claude Desktop **Find your Claude Desktop config file:** - **macOS:** `~/Library/Application Support/Claude/claude_desktop_config.json` - **Windows:** `%APPDATA%\Claude\claude_desktop_config.json` - **Linux:** `~/.config/Claude/claude_desktop_config.json` **Add this configuration (replace `/path/to/your/projects` with your actual path):** ```json { "mcpServers": { "xplainable": { "command": "/path/to/your/projects/xplainable-mcp-server/xplainable-mcp-env/bin/python", "args": ["-m", "xplainable_mcp.server"], "env": { "XPLAINABLE_API_KEY": "your-api-key-here", "XPLAINABLE_HOST": "https://platform.xplainable.io" } } } } ``` **For this specific setup, the configuration would be:** ```json { "mcpServers": { "xplainable": { "command": "/Users/jtuppack/projects/xplainable-mcp-server/xplainable-mcp-env/bin/python", "args": ["-m", "xplainable_mcp.server"], "env": { "XPLAINABLE_API_KEY": "your-api-key-here", "XPLAINABLE_HOST": "https://platform.xplainable.io" } } } } ``` **If you already have other MCP servers configured,** just add the `"xplainable"` section inside your existing `"mcpServers"` object: ```json { "mcpServers": { "existing-server": { "command": "some-command" }, "xplainable": { "command": "/Users/jtuppack/projects/xplainable-mcp-server/xplainable-mcp-env/bin/python", "args": ["-m", "xplainable_mcp.server"], "env": { "XPLAINABLE_API_KEY": "your-api-key-here", "XPLAINABLE_HOST": "https://platform.xplainable.io" } } } } ``` ### 4. Restart Claude Desktop Close and reopen Claude Desktop. The Xplainable tools should now be available! ## What You Can Do Once configured, you can ask Claude to: - List your team's models: *"Show me all the models in my team"* - Get model details: *"Get information about model ID xyz"* - List deployments: *"What deployments do I have?"* - Deploy a model: *"Deploy version abc123 of my model"* - Generate deployment keys: *"Create a deployment key for deployment xyz"* - Get model reports: *"Generate a report for my model"* ## Troubleshooting **"MCP server not found" error:** - Verify the Python path in your config is correct - Make sure the virtual environment exists and has the MCP server installed - Check that your API key is correct - Restart Claude Desktop after making config changes **"Module not found" error:** - Activate your virtual environment: `source xplainable-mcp-env/bin/activate` - Reinstall the server: `pip install -e . --force-reinstall` - Verify installation: `python -m xplainable_mcp.server --help` **Connection errors:** - Verify your API key at https://platform.xplainable.io - Make sure `XPLAINABLE_HOST` is set correctly - Check your internet connection **Permission errors:** - Some operations require special permissions in your Xplainable account - Contact your team admin if you can't perform certain actions **Virtual environment issues:** - Make sure you're using Python 3.9 or later - Recreate the environment if needed: `rm -rf xplainable-mcp-env && python3 -m venv xplainable-mcp-env` - Check that all dependencies installed correctly ## Advanced Configuration **Enable write operations (deployment, keys, etc.):** ```json "env": { "XPLAINABLE_API_KEY": "your-api-key-here", "XPLAINABLE_HOST": "https://platform.xplainable.io", "ENABLE_WRITE_TOOLS": "true" } ``` **Connect to a different Xplainable instance:** ```json "env": { "XPLAINABLE_API_KEY": "your-api-key-here", "XPLAINABLE_HOST": "https://your-custom-instance.com" } ``` **Update the MCP server:** ```bash cd /path/to/your/projects/xplainable-mcp-server source xplainable-mcp-env/bin/activate git pull origin main pip install -e . --force-reinstall ``` **Check available tools:** ```bash source xplainable-mcp-env/bin/activate python -c " from xplainable_mcp.server import list_tools tools = list_tools() print(f'Total tools: {tools[\"total_tools\"]}') for category, tool_list in tools['categories'].items(): print(f'{category}: {len(tool_list)} tools') " ``` --- **Need help?** Open an issue at: https://github.com/xplainable/xplainable-mcp-server/issues