A Model Context Protocol (MCP) server that provides comprehensive SQLite database management and analysis capabilities. This server allows LLMs to explore database schemas, query data, perform updates, and conduct statistical analysis.

• Schema Exploration

  • List all tables in the database
  • View detailed schema information for specific tables
  • Examine column types and constraints

• Data Management

  • Execute read-only SQL queries
  • Perform data modifications (UPDATE, INSERT, DELETE)
  • Safe execution with error handling

• Data Analysis

  • Basic statistical analysis (row counts, null counts, numeric stats)
  • Detailed analysis including categorical data distributions
  • Automatic type detection and appropriate statistical measures

• Python 3.8 or higher
• SQLite database file
• Claude Desktop (optional, for desktop integration)

1. First, ensure you have the required Python packages:

pip install mcp pandas

2. Download the SQLite MCP server script:

# Clone this repository or download sqlite_mcp.py directly
curl -O https://raw.githubusercontent.com/yourusername/sqlite-mcp/main/sqlite_mcp.py

3. For Claude Desktop integration:

# Install using MCP CLI
mcp install sqlite_mcp.py --name "SQLite Explorer" --env DB_PATH=/path/to/your/database.sqlite

• Locate the claude_desktop_config.json file and add below to the mcpServers section
• change the paths to the correct ones for your system.
• Set database location in DB_PATH variable in the .env file.

"sqlite_mcp": {
    "command": "C:\\path\\to\\python.exe",
    "args": [
    "C:\\path\\to\\sqlite-mcp\\server.py"
    ]
}

The server exposes the following MCP resources:

• schema://tables

  • Lists all available tables in the database
  • Example response:

    Available tables:
    - users
    - products
    - orders
    

• schema://{table}

  • Returns detailed schema information for a specific table
  • Example response:

    Table: users
    
    Create Statement:
    CREATE TABLE users (
        id INTEGER PRIMARY KEY,
        name TEXT NOT NULL,
        email TEXT UNIQUE
    )
    
    Columns:
    - id (INTEGER) NOT NULL PRIMARY KEY
    - name (TEXT) NOT NULL
    - email (TEXT)
    

Execute read-only SQL queries:

SELECT * FROM users LIMIT 5

Perform data modifications:

INSERT INTO users (name, email) VALUES ('John Doe', 'john@example.com')

UPDATE users SET email = 'new@example.com' WHERE id = 1

Perform statistical analysis on table data:

Parameters:

• table: Name of the table to analyze
• analysis_type: Either 'basic' or 'detailed'

Example response:

{
  "row_count": 1000,
  "column_count": 5,
  "null_counts": {
    "id": 0,
    "name": 0,
    "email": 15
  },
  "numeric_columns": {
    "id": {
      "mean": 500.5,
      "std": 288.819,
      "min": 1,
      "max": 1000
    }
  }
}

The server implements several security measures:

1. Input validation for all SQL operations
2. Read-only queries are separated from data modifications
3. Database connection error handling
4. SQL injection protection through parameterized queries

The server provides clear error messages for common issues:

• Database connection failures
• Invalid SQL syntax
• Table not found errors
• Permission issues
• Type mismatches

Contributions are welcome! Please feel free to submit a Pull Request.

This project is licensed under the MIT License - see the LICENSE file for details.