SQL Connector MCP

An MCP (Model Context Protocol) server that provides SQL database connectivity for AI assistants like Claude.

Features

• Database Exploration: List all table names or filter tables by substring
• Schema Inspection: Get detailed schema definitions including columns, types, primary keys, and foreign key relationships
• Query Execution: Execute SQL queries with parameterized inputs and formatted results
• Form Fetching: Retrieve form schemas from database based on trigger keywords

Installation

pip install sql-connector-mcp

Configuration

Set the following environment variables:

# Required: Database connection URL
DB_URL=mysql://user:password@host:port/database

# Optional: Maximum characters for query results (default: 4000)
EXECUTE_QUERY_MAX_CHARS=4000

# Optional: Path for saving full result sets
CLAUDE_LOCAL_FILES_PATH=/path/to/files

Usage

Run the MCP server:

python -m sql_connector_mcp.server

Opens the caldue code or cursor or any ai ide and add this json "my_database": { "command": "python", "args": ["-m", "sql_connector_mcp.server"], "env": { "DB_URL": "your dbatabase url" }

Available Tools

Requirements

• Python >= 3.7
• SQLAlchemy
• PyMySQL
• FastMCP

License

MIT License

Author

Battula Anil Kumar (anilkumarbaatula@gmail.com)