MCP SQLite Tool

An MCP server that provides AI with secure and direct access to SQLite databases.

🌟 Features

• Secure Access: Allows AI to interact with SQLite databases while validating file paths to prevent unauthorized access.
• Single-File Server: The core logic is contained in a single Python file, making it easy to understand and maintain.
• Standard I/O (stdio) Transport: Implements the official MCP protocol using standard I/O for seamless integration with tools like VS Code.
• Pythonic & Easy to Use: Built with the fastmcp framework, which uses decorators to simplify tool creation.

🚀 Installation

This tool is available as a package on PyPI. You can install it using pip.

pip install mcp-sqlite-tool

Note: This tool requires Python 3.10 or higher. The mcp[cli] dependency will be installed automatically.

🤖 Usage with VS Code

To use this tool with an MCP-compatible client like VS Code, you need to configure your workspace to discover and run the server.

1. Install the package: Ensure you have the mcp-sqlite-tool package installed in the Python environment that VS Code is using.

   pip install mcp-sqlite-tool
   

2. Configure settings.json: Open the Command Palette in VS Code (Ctrl+Shift+P), search for "Open Workspace Settings (JSON)", and add the following configuration:

   {
    "mcpServers": {
     "sqlite-query": {
      "command": "mcp-sqlite-tool"
     }
    }
   }
   

   • "sqlite-query": This name identifies the server. It must match the name specified in the Python tool's code (FastMCP("sqlite-query")).
   • "command": "mcp-sqlite-tool": This is the command-line entry point installed by pip. VS Code will use this to launch your server.

3. Interact with the AI: Once configured, the AI client will be able to use the tool execute_sqlite_query. You can now prompt the AI to interact with your SQLite files.

Example Prompt

"Using the sqlite-query tool, tell me the names and emails of all users in the users table of C:\Users\John\Documents\project.db."

The AI will interpret this request and call your tool with the appropriate arguments.

🛠️ Tool Documentation

The core tool provided by this server is:

execute_sqlite_query

Executes a given SQL query on a specified SQLite database file.

Returns: A dictionary containing the query results or an error message if the operation fails.

🤝 Contributing

Contributions, issues, and feature requests are welcome! If you have any issues, please contact me directly on Telegram (@mrbeandev).

📄 License

Distributed under the MIT License. See LICENSE for more information.