MCP Server for Paddle Billing

Paddle Billing is the developer-first merchant of record. We take care of payments, tax, subscriptions, and metrics with one unified API that does it all.

This is a Model Context Protocol (MCP) server that provides LLMs and AI agents with tools for interacting with the Paddle API.

Important: This MCP server works with Paddle Billing. It does not support Paddle Classic. To work with Paddle Classic, see: Paddle Classic API reference

Features

The MCP server has near parity with the Paddle API, allowing AI assistants and agents to:

• Manage your full Paddle catalog
• View customer, purchase, and provisioning information
• Handle subscription, payment, and refund workflows
• Debug billing and order management issues
• Create and adjust transactions directly in conversation
• Generate financial reports for financial and operational insights
• Implement and test Paddle integrations faster

Available tools

The MCP server can use the following tools to take actions with your Paddle account:

Development

1. Install dependencies:

   Copy

   pnpm install

2. Build the server:

   Copy

   pnpm build

3. Update client to use the local build:

   Copy

   {
     "mcpServers": {
       "paddle": {
         "command": "node",
         "args": ["path/to/paddle-mcp-server/build/index.js"],
         "env": {
           "PADDLE_API_KEY": "your_api_key",
           "PADDLE_ENVIRONMENT": "sandbox",
           "PADDLE_MCP_TOOLS": "all"
         }
       }
     }
   }

   The PADDLE_MCP_TOOLS environment variable accepts the same values as the --tools argument: all, read-only, non-destructive, or a comma-separated list of tool names. If not set, defaults to non-destructive.

Debugging

To debug the MCP server, you can use the MCP Inspector tool:

1. Run the server with the inspector:

   Copy

   pnpm inspector

2. Open the provided URL in your browser to view and debug the MCP requests and responses.

3. Include the --api-key and --environment arguments.

Learn more

• Paddle developer docs
• Paddle API reference
• Sign up for Paddle Billing

Installation

To use the MCP server, you'll need an API key. You can create and manage API keys in Paddle > Developer tools > Authentication:

• Sandbox: https://sandbox-vendors.paddle.com/authentication-v2
• Live: https://vendors.paddle.com/authentication-v2

Adding the following to your MCP settings file will configure and run the Paddle MCP server in a client like Claude Desktop, Cursor or Windsurf:

Method 1: One-click installation in Cursor or VS Code

You can install the Paddle MCP server with a single click in Cursor or VS Code.

After installation, you'll need to update the configuration in your MCP settings file to replace your_api_key with your actual Paddle API key and adjust the environment and tools values as needed.

Method 2: Using command-line arguments (Recommended for npx)

Add the following to the MCP settings or configuration file in the client you're using:

{
  "mcpServers": {
    "paddle": {
      "command": "npx",
      "args": [
        "-y",
        "@paddle/paddle-mcp",
        "--api-key=your_api_key",
        "--environment=sandbox",
        "--tools=non-destructive"
      ]
    }
  }
}

Replace your_api_key with your actual Paddle API key, set --environment to either sandbox or production, and set --tools to the tools which you want to be loaded and available to the MCP client.

Method 3: Using environment variables

Add the following to the MCP settings or configuration file in the client you're using:

{
  "mcpServers": {
    "paddle": {
      "command": "npx",
      "args": ["-y", "@paddle/paddle-mcp"],
      "env": {
        "PADDLE_API_KEY": "your_api_key",
        "PADDLE_ENVIRONMENT": "sandbox",
        "PADDLE_MCP_TOOLS": "non-destructive"
      }
    }
  }
}

Replace your_api_key with your actual Paddle API key, set PADDLE_ENVIRONMENT to either sandbox or production, and set PADDLE_MCP_TOOLS to the tools which you want to be loaded and available to the MCP client.

Filtering tools

You can filter which tools are loaded and available to the MCP client by passing the --tools argument (Method 2) or setting the PADDLE_MCP_TOOLS environment variable (Method 1 or 3). Accepted values are:

• all - All tools are available
• read-only - Only read operations are available
• non-destructive - Read operations and safe write operations are available (default)
• A comma-separated list of specific tool names (e.g., list_products,get_product,create_product)

For detailed setup guides, see:

• Claude Desktop
• Cursor
• Windsurf