MySQL MCP Server (Python & FastAPI)

A powerful Model Context Protocol (MCP) server that connects your AI assistant (like Claude Desktop) directly to your MySQL database. This server is written in Python, uses aiomysql for asynchronous database pooling, and is powered by FastAPI to support both standard STDIO and HTTP-based SSE transports.

🚀 Features

This server implements a wide range of MCP capabilities to give you full visibility and control over your MySQL database.

🛠️ Tools

Table Management

• list-tables: List all tables in the database.

• describe-table: Get the schema (columns, types, keys) of a specific table.

• create-table: Create a new table using a SQL query.

Data Operations (CRUD)

• insert-row: Insert a new row into a table using JSON data.

• update-row: Update existing rows with a convenient JSON-based WHERE clause.

• delete-row: Delete rows with a safety-first JSON-based WHERE clause.

Query Execution

• execute-query: Execute read-only SQL queries (SELECT, SHOW, DESCRIBE). Safe for exploration.

• execute-write-query: Execute write SQL queries (INSERT, UPDATE, DELETE, ALTER, etc.). Use with caution.

Database Administration

• create-database: Create a new database.

Advanced Objects Inspection

• Stored Procedures: list-procedures, show-procedure (view CREATE statement).

• Functions: list-functions, show-function (view CREATE statement).

• Triggers: list-triggers, show-trigger.

• Views: list-views, show-view.

📦 Resources

Access database content directly as standard MCP resources:

• mysql://<table_name>/schema: View the full JSON schema of a table.

• mysql://<table_name>/data: View the first 100 rows of a table as JSON.

💬 Prompts

Pre-built templates to help the AI perform complex tasks:

• analyze-table: "Analyze the structure and sample data for table 'Users'..."

  • Automatically fetches schema and sample data to provide insights.

• optimize-query: "Analyze this query for performance..."

  • Helps you write faster, more efficient SQL.

Related MCP server: MCPDB - Database Access MCP Server

📦 Installation & Setup

1. Clone the repository:

   git clone <repository-url>
   cd mysql-mcp-server

2. Install the dependencies: Ensure you have Python 3.10+ installed. Install requirements using pip:

   pip install -r requirements.txt

⚙️ Running the Server

The server can run in two modes:

Mode 1: STDIO Transport (Local Integration)

Directly run the script with your database URL as the argument. This is the mode used by desktop clients like Claude Desktop.

python main.py mysql://user:password@localhost:3306/database_name

Mode 2: SSE Transport (Web / Networked Server via FastAPI)

Run the FastAPI web server by specifying a port:

python main.py --port 8000 mysql://user:password@localhost:3306/database_name

Alternatively, you can set the environment variable and run via standard uvicorn:

# On Windows (cmd):
set MYSQL_URL=mysql://user:password@localhost:3306/database_name
# On Windows (PowerShell):
$env:MYSQL_URL="mysql://user:password@localhost:3306/database_name"
# On Linux/macOS:
export MYSQL_URL="mysql://user:password@localhost:3306/database_name"

uvicorn main:app --host 0.0.0.0 --port 8000

Visit http://localhost:8000 to view the server status and connection dashboard.

🛠️ Claude Desktop Configuration

Add the following block to your claude_desktop_config.json depending on your preferred transport:

Option A: Direct STDIO Connection (No server to run in background)

{
  "mcpServers": {
    "mysql-stdio": {
      "command": "python",
      "args": [
        "c:/Users/bhave/OneDrive/Desktop/MySQL MCP Server/main.py",
        "mysql://user:password@localhost:3306/database_name"
      ]
    }
  }
}

Note: Ensure you use the absolute path to main.py on your machine.

Option B: Remote SSE Connection (Requires running FastAPI server on port 8000)

{
  "mcpServers": {
    "mysql-sse": {
      "command": "python",
      "args": [
        "-m",
        "mcp.client.sse",
        "http://localhost:8000/sse"
      ]
    }
  }
}

💡 Usage Examples

Once connected in Claude, you can use queries like:

• Exploration: "What tables are in my database?"

• Schema: "Show me the schema for the orders table."

• Data Analysis: "Select the top 5 customers who spent the most money."

• Modification: "Insert a new user named 'Alice' into the users table."

• Optimization: "Analyze this query and tell me why it's slow: SELECT * FROM logs"

🔒 Security Note

This server has powerful capabilities, including write access (execute-write-query, delete-row, etc.). Always ensure you are connecting to a database where you have appropriate permissions, and be careful when asking the AI to modify data.

📄 License

MIT