XiYan

Install xiyan-mcp-server into your Python environment. Python 3.11+ is required.

pip install xiyan-mcp-server

Copy the demo config and fill in your LLM credentials and database connection details. The model name, key, and url select which text-to-SQL backend to use.

cat config_demo.yml

Edit the config file with your actual values. Using the Modelscope-hosted XiYanSQL-QwenCoder-32B model gives the best accuracy. Set database.host, port, user, password, and database to match your setup.

mcp:
  transport: "stdio"
model:
  name: "XGenerationLab/XiYanSQL-QwenCoder-32B-2412"
  key: "YOUR_MODELSCOPE_API_KEY"
  url: "https://api-inference.modelscope.cn/v1/"
database:
  host: "localhost"
  port: 3306
  user: "root"
  password: "your_password"
  database: "your_database_name"

Open ~/Library/Application Support/Claude/claude_desktop_config.json and add a mcpServers entry that launches the server with your config file.

{
  "mcpServers": {
    "xiyan-mcp": {
      "command": "python",
      "args": ["-m", "xiyan_mcp_server", "--config", "/path/to/your/config.yml"]
    }
  }
}

The server exposes your database tables as MCP resources using the pattern mysql://table_name or postgresql://table_name. Ask your AI assistant to list available tables to confirm the connection is working.

Use the get_data tool by asking questions in plain language. The server translates your question to SQL using XiYanSQL, executes it, and returns results.