🐬 MySQL MCP Server (FastMCP-Powered)

Leverage the power of AI to automate your MySQL workflows with FastMCP.
This tool enables AI agents like Claude, OpenAI, or DeepSeek to interact with your MySQL database using natural language, making database management seamless and intelligent.

Whether you're generating synthetic data, querying your tables, or performing updates, this MCP server allows AI to operate on your database through standardized interfaces — all from your local machine.

🔎 Best use for generating synthetic data for testing your apis.

💡 Key Features

• Automates SQL tasks via AI (Claude, OpenAI, DeepSeek, etc.)
• Works over Stdio, ideal for local desktop integration
• Generates synthetic data based on your schema and recent values
• No manual SQL knowledge required
• Highly useful for test environments and data prototyping

---

🛠️ Supported Tools

Tool	Description
get_column_names(table_name)	Retrieve all column names of a table
get_all_table_names()	List all table names in the database
select(...)	Fetch records from any table
insert(...)	Insert data into a table
update(...)	Update records with conditions
delete(...)	Delete records based on criteria

These are internally used by the AI to understand and manipulate your database efficiently.

---

⚙️ Configuration & Setup

📋 Prerequisites

• Python 3.10+
• MySQL (running locally or remotely)
• Anthropic Claude Desktop app (or Cursor)
• UV (Python package manager), install with curl -LsSf https://astral.sh/uv/install.sh | sh

Steps

1. Clone this repository

   git clone https://github.com/Ashutosh-rajput/mysql-mcp.git
   cd mysql-mcp
   uv pip install -r requirements.txt
   

2. Fill the .env

   MYSQL_HOST=localhost
   MYSQL_PORT=3306
   MYSQL_USER=root
   MYSQL_PASSWORD=yourpassword
   MYSQL_DATABASE=yourdatabase
   

3. Connect to the MCP server

   Copy the below json with the appropriate {{PATH}} values:

       {
      "mcpServers": {
        "Mysql": {
          "command": "[path to uv]/uv",
          "args": [
            "run",
            "--with", "fastmcp",
            "--with", "mysql-connector-python",
            "--with", "python-dotenv",
            "fastmcp",
            "run",
            "[path to your main.py]/main.py"
          ]
        }
      }
    }
   

   For Claude, save this as claude_desktop_config.json in your Claude Desktop configuration directory at:

   ~/Library/Application Support/Claude/claude_desktop_config.json
   

   For Cursor, save this as mcp.json in your Cursor configuration directory at:

   ~/.cursor/mcp.json
   

4. Restart Claude Desktop / Cursor

   Open Claude Desktop and you should now see WhatsApp as an available integration.

   Or restart Cursor.

🧠 Tip

Use this project as a backbone to build intelligent, event-driven Mysql bots, agents.

---

📜 License

MIT License. Feel free to use and modify as per your needs.

---

👤 Author: Ashutosh Rajput
🌐 Project: MySql-MCP