Turn Any API Into An MCP Server

How to Connect Any API to ChatGPT and Claude Using MCP

APIs are the backbone of modern software development, powering everything from e-commerce platforms to data analytics tools. But what if you could interact with these APIs using natural language instead of writing code? Thanks to the Model Context Protocol (MCP), you can now ask questions like "How many orders came in the last 24 hours in these four states?" and have your AI assistant handle the rest.

In this guide, we'll walk through how to transform any REST API into an MCP server that works seamlessly with LLMs like ChatGPT, Claude, and Cursor.

What is MCP and Why Does It Matter?

The Model Context Protocol (MCP) is a game-changing standard that allows large language models to interact directly with external tools and APIs. Instead of manually crafting API queries, you can simply describe what you need in plain English, and the LLM will:

1. Understand your request

2. Call the appropriate API endpoint

3. Process the response

4. Reply to you in natural language

This creates a bridge between conversational AI and the programmatic world of APIs, making complex data queries accessible to anyone.

Introducing OpenAPI to MCP Converter

The tool that makes this magic possible is **OpenAPI 2 MCP**, a free converter available on Hugging Face Spaces. This tool takes any API with an OpenAPI specification and transforms it into an MCP server that your favorite AI assistants can use.

 Getting Started

Here's how to convert your API in just a few steps:

**Step 1: Sign In**

Navigate to the OpenAPI 2 MCP tool on Hugging Face Spaces and sign in with your Hugging Face account. The tool is completely free to use, though you'll need to authorize certain permissions.

**Step 2: Provide Your API Specification**

OpenAPI is the standard format for describing REST APIs. It includes information about all available endpoints, what data they expect, and what they return. You'll need to provide two pieces of information:

- The URL to your OpenAPI JSON specification

- The base URL where API requests should be sent (for example, `yourserver.com/api/v1`)

**Step 3: Filter Endpoints (Optional)**

Once you input your API spec, the tool will discover all available endpoints. For example, a pet store API might reveal 19 different endpoints covering pets, store inventory, and users. If you only need specific functionality, you can filter the endpoints. Selecting just "pets" might narrow it down to eight relevant endpoints.

**Step 4: Generate Your MCP Space**

After configuration, the tool creates a new MCP-enabled space. This space will take a moment to build and start, but once it's ready, you'll have a fully functional MCP server.

Testing and Using Your MCP Server

Your new MCP space comes with several useful features:

**Manual Testing**

You can test individual endpoints directly in the interface. For instance, querying for "pet ID 1" might return information about a dog with an "available" status. This is helpful for verifying that your API connections are working correctly.

**MCP Configuration**

The real power comes from the MCP configuration that's automatically generated. This config file can be copied and pasted directly into your AI tool of choice. 

 Connecting to Cursor (or Other LLMs)

Let's look at how to connect your new MCP server to Cursor as an example:

1. Open Cursor and navigate to Settings > Tools > MCP Tools

2. Paste your MCP configuration

3. The tools will appear as enabled

Now you can ask natural language questions like "What kind of pet is pet one?" The LLM will automatically:

- Identify which tool to use

- Call the appropriate endpoint

- Parse the response

- Answer in plain English ("Pet one is a dog and its status is available")

The Power of Automatic Tool Discovery

What makes this approach particularly powerful is that the MCP converter automatically extracts rich information from your API specification:

- Detailed descriptions of what each endpoint does

- Required and optional parameters

- Available choices for enumerated fields

- Default values

- Expected response formats

For example, an endpoint that retrieves pets by status will automatically understand the valid status options (available, pending, sold) and present these choices to the LLM.

Practical Applications

This technology opens up numerous possibilities:

**E-commerce Management**: Ask about recent orders, inventory levels, or customer data without writing SQL queries or API calls.

**Data Analytics**: Request complex reports by simply describing what you want to know.

**System Administration**: Check server status, user accounts, or system metrics through conversation.

**Content Management**: Query and update your CMS by describing changes in natural language.

 Getting Started with Your Own APIs

The process works with any RESTful API that has an OpenAPI specification. Whether you're working with internal company APIs, third-party services, or your own custom endpoints, the conversion process remains the same.

The beauty of this approach is its generality. Once you understand the pattern of converting an API spec to an MCP server, you can apply it to any API in your toolkit. This means all your programmatic tools can become conversational interfaces accessible to anyone on your team, regardless of technical expertise.

Conclusion

The Model Context Protocol represents a significant leap forward in making technical tools accessible through natural language. By converting your APIs to MCP servers, you're not just creating another interface—you're fundamentally changing how people can interact with your data and services.

The OpenAPI 2 MCP converter makes this transformation straightforward and free. In just a few minutes, you can take any existing API and make it conversable, opening up new possibilities for how your team works with data and services.

Ready to get started? Find the OpenAPI 2 MCP tool on Hugging Face Spaces and start making your APIs conversational today.