> For the complete documentation index, see [llms.txt](https://developer.paddle.com/llms.txt).

# Interact with Paddle using AI

Use AI assistants like Claude or AI-powered IDEs like Cursor to create your product catalog, manage your billing and subscriptions, and generate reports.

---

{% callout type="info" %}
AI tools using the MCP server can access your Paddle data. Only use tools or assistants that you trust.
{% /callout %}

## What's new?

We've released a [Model Context Protocol (MCP) server](https://modelcontextprotocol.io/introduction) that you can use to interact with the Paddle platform using AI assistants. This means you can work with Paddle Billing using natural language, speeding up integration and making it easier to get access to your data.

Using [our MCP server](https://github.com/PaddleHQ/paddle-mcp-server/), AI assistants can help you create products and prices, analyze subscription data, and generate reports — all through chat, with no coding required.

## How it works

Model Context Protocol (MCP) servers are a way for you to connect AI tools, like [Claude](https://claude.ai/) or [Cursor](https://www.cursor.com/), directly to platforms like Supabase, Vercel, and Cloudflare. Rather than making requests using an API or using an interface like the dashboard, you can work with your business data using natural language chat.

Now, we've released [an MCP server for Paddle Billing](https://github.com/PaddleHQ/paddle-mcp-server/). You can use it to streamline workflows and quickly access Paddle functionality without context switching or writing custom code.

This release includes support for:

* Getting information about [products](https://developer.paddle.com/api-reference/products/overview.md) and [prices](https://developer.paddle.com/api-reference/prices/overview.md)
* Creating new products and prices
* Getting information about [customers](https://developer.paddle.com/api-reference/customers/overview.md)
* Getting information about [subscriptions](https://developer.paddle.com/api-reference/subscriptions/overview.md) and [transactions](https://developer.paddle.com/api-reference/transactions/overview.md)
* Creating [reports](https://developer.paddle.com/api-reference/reports/overview.md)

You can connect multiple MCP servers with AI assistants to create complex testing and integration workflows across platforms. For example, you can use MCP servers in Cursor to check data across Paddle, Supabase, and other platforms in your stack to help you troubleshoot problems.

## Examples

{% callout type="info" %}
You might get different responses as AI assistants can be unpredictable. Always check created data and responses carefully.
{% /callout %}

* [**"Get information about subscriptions for `parker@example.com`"**](https://claude.ai/share/621bb420-bf1f-49e4-8790-7e20d5bcab71)

   In this example, Claude fetches information about a customer and their related subscriptions and transactions to generate a quick visual summary. This workflow is great when you're on-the-go and want to dig into your data.

* [**"Model prices for me"**](https://claude.ai/share/f7cd1bff-1dd2-40a9-afc1-88a7e4f851c1)

   In this example, Claude creates a three-tier pricing structure with monthly and annual plans, including localized prices for the US, the UK, Germany, Brazil, and India. This workflow is especially powerful in IDEs like [Cursor](https://www.cursor.com/) using your pricing page as context.

## Next steps

The Paddle MCP server is available now.

Setup instructions vary depending on the AI assistant that you're using. For more details, see [@PaddleHQ/paddle-mcp-server on GitHub](https://github.com/PaddleHQ/paddle-mcp-server).