mcp_sqlite_poc

mcp_sqlite_poc

0
databases

The SQLite MCP Server is a Model Context Protocol server facilitating interaction between AI models and SQLite databases using standardized tools. It offers features like query execution, schema discovery, and feedback collection and supports both read-only and write operations with security considerations.

SQLite MCP Server

A Model Context Protocol (MCP) server implementation for SQLite databases, allowing AI models to interact with SQLite databases through standardized tools.

Features

  • Query Execution: Execute SQL queries with built-in validation and safety checks
  • Schema Discovery: List tables and describe table schemas
  • Row Counting: Count rows in specified tables
  • Sample Data: Insert sample data for demonstration purposes (when not in read-only mode)
  • Automatic Initialization: Creates a sample database with products, customers, orders, and feedback if none exists
  • Read-Only Mode: Option to restrict to read-only operations for safety
  • Feedback Collection: Tool to collect user feedback and store it in the database

Requirements

  • Python 3.8+
  • Dependencies listed in requirements.txt

Installation

  1. Clone this repository
  2. Install dependencies:
pip install -r requirements.txt
  1. Configure environment variables (optional):
    • Copy .env.example to .env (or use the existing .env file)
    • Modify settings as needed

Usage

Starting the Server

python server.py

The server will start on port 8080 (or the port specified in your .env file).

Configuration Options

Configure the server by editing the .env file:

DB_PATH=sample.db         # Path to your SQLite database file
MCP_PORT=8080             # Port to run the MCP server on
READ_ONLY=true            # Set to false to allow write operations

Using with Claude Desktop

Add this server to your Claude Desktop configuration:

{
"mcpServers": {
		"sqlite": {
			"command": "<Path to UV>",
			"args": [
				"--directory",
				"<Absolute Path to Project Workspace>",
				"run",
				"server.py"
			]
		}
	}
}

Available Tools

The MCP server provides the following tools:

  1. execute_query - Execute SQL queries on the database
  2. list_tables - List all tables in the database
  3. describe_table - Get the schema for a specific table
  4. count_rows - Count rows in a specific table
  5. insert_sample_data - Insert sample data (disabled in read-only mode)
  6. add_feedback - Add user feedback to the feedback table (disabled in read-only mode)

Sample Database Schema

The sample database includes:

Products Table

  • id (INTEGER, PRIMARY KEY)
  • name (TEXT, NOT NULL)
  • description (TEXT)
  • price (REAL, NOT NULL)
  • category (TEXT)
  • in_stock (BOOLEAN)

Customers Table

  • id (INTEGER, PRIMARY KEY)
  • name (TEXT, NOT NULL)
  • email (TEXT, UNIQUE)
  • signup_date (TEXT)

Orders Table

  • id (INTEGER, PRIMARY KEY)
  • customer_id (INTEGER, FOREIGN KEY)
  • order_date (TEXT)
  • total_amount (REAL)

Feedback Table

  • id (INTEGER, PRIMARY KEY)
  • user (TEXT, NOT NULL)
  • email (TEXT, NOT NULL)
  • feedback (TEXT, NOT NULL)
  • submission_date (TEXT, DEFAULT CURRENT_TIMESTAMP)

Security Considerations

  • SQL queries are validated to prevent dangerous operations
  • In read-only mode, all data modification operations are blocked
  • Consider running in read-only mode for production use

License

MIT License