SQLite MCP Server

This is a Model Context Protocol (MCP) server that provides access to a SQLite database. It allows AI assistants (like Trae) to query and modify SQLite databases directly.

Features

• Read Queries: Execute SELECT queries to retrieve data.
• Write Queries: Execute INSERT, UPDATE, DELETE queries to modify data.
• Schema Inspection: List tables and describe table schemas.
• Secure: Runs locally on your machine.

Prerequisites

• Python 3.10 or higher
• pip (Python package installer)

Installation

1. Clone or download this repository.
2. Install the required dependencies:

pip install -r requirements.txt

Usage

You can run the server directly from the command line:

python main.py --db path/to/your/database.sqlite

If the database file does not exist, it will be created automatically when you perform a write operation.

Configuration in Trae

To use this MCP server in Trae, you need to add it to your MCP configuration file.

1. Open Trae.
2. Go to Settings -> MCP Servers (or edit the configuration file directly if you know the location, typically ~/.config/trae/config.json or similar depending on OS).
3. Add the following configuration:

{
  "mcpServers": {
    "sqlite": {
      "command": "python",
      "args": [
        "absolute/path/to/sqlite_mcp/main.py",
        "--db",
        "absolute/path/to/your/database.sqlite"
      ]
    }
  }
}

Note:

• Replace absolute/path/to/sqlite_mcp/main.py with the full path to the main.py file in this project.
• Replace absolute/path/to/your/database.sqlite with the full path to your SQLite database file.
• On Windows, use double backslashes \\ or forward slashes / in paths (e.g., "C:\\Users\\Name\\sqlite_mcp\\main.py").

API Documentation

Tools

read_query

Executes a SELECT query on the SQLite database.

• Input: query (string) - The SQL SELECT query.
• Output: List of dictionaries representing the rows.

write_query

Executes an INSERT, UPDATE, or DELETE query.

• Input: query (string) - The SQL modification query.
• Output: Success message with row count.

list_tables

Lists all tables in the database.

• Input: None.
• Output: List of table names.

describe_table

Gets the schema for a specific table.

• Input: table_name (string).
• Output: List of column definitions.