Connect AI assistants to your Paddle account
Our Model Context Protocol (MCP) server has been updated with 80+ tools to enable powerful new workflows within AI assistants and agents.
What's new?
We've significantly expanded our Model Context Protocol (MCP) server to include the full suite of Paddle API functionality.
You can now use AI assistants to handle a variety of new AI-powered use-cases and workflows, including building and testing your integration, debugging complex billing issues, processing enterprise deals, and managing subscription lifecycle changes.
How it works
The Paddle Model Context Protocol (MCP) server enables AI systems to interact with your Paddle account. It acts as a "universal remote" that exposes data and actions to Large Language Models (LLMs), so they're empowered to do more than only chat.
Previously, our MCP server supported a subset of functionality in the Paddle API. With this update, we added full support for the Paddle API. This means AI agents have access to over 80 tools covering every aspect of Paddle Billing, letting you and your team interact with Paddle fully using natural language.
Registering tools
By default, only non-destructive tools are available. This means that the AI assistant can only use tools that don't edit or delete any records.
You can expand or restrict the tools that are available to the AI assistant by setting the tools flag or environment variable when installing or configuring the MCP server.
| Value | Description |
|---|---|
all | Load all tools into the AI assistant. |
read-only | Load read-only tools (fetch and list operations) into the AI assistant. |
non-destructive | Load tools that don't edit or delete any records into the AI assistant. Default. |
| Comma-separated list of tool names | Load only the specified tools (e.g. create_product,create_price,list_products,list_prices) into the AI assistant. |
Examples
Building and testing integrations
Integrate Paddle, or update your existing integration, by using the MCP server in your AI-powered IDE. Configure webhook endpoints, create notification settings, set up test simulations, debug webhook delivery issues, and replay any failed events while testing.
Create and update products, prices, discounts, and discount groups. Handle more complex workflows, like pricing changes, grandfathering, and customer pricing migrations.
Debugging billing issues end-to-end
Investigate and identify bugs related to your billing setup, explain complex billing scenarios, and issue refunds or credits if appropriate. Replay webhooks to fix provisioning issues when your webhook endpoint server has been down.
Set up enterprise customers with custom pricing, manual invoicing, and specific payment terms. Use in conjunction with quote management tools and CRMs. Migrate self-serve customers to enterprise plans with proper proration and generate invoice PDFs immediately.
Connect subscription data with customer history, usage patterns, and business policies from other sources, enabling use cases like identifying at-risk trials and generating targeted retention offers. Preview financial impacts before executing upgrades, downgrades, pauses, cancellations, and adjustments.
Next steps
The updates to the Paddle MCP server are available now.
You can install the MCP server on any clients which support MCP servers, like Claude Desktop and Cursor. No action is needed to gain access to all non-destructive tools if you installed or configured the MCP server in the past.
If you want to further limit or expand the tools that are available to the AI assistant, set the tools parameter.
Explore all the tools, capabilities, best practices, and example use cases by reading our guide on the Paddle MCP server.