OpenAPI MCP server

This is a small helper program that reads an API's instruction manual (called an OpenAPI or Swagger file) and turns it into actions your AI assistant can perform. It is handy if you want your AI to talk to a service that already publishes one of these manuals, without anyone writing new code. You run it yourself, usually with Docker, and point it at the manual.

What this MCP server does

You give this helper a file or web link that describes an API, and it reads that description. From it, the helper builds a set of actions your AI can call, one for each thing the API can do. When you ask your AI something, the AI picks the right action, the helper sends the request to the real API, and the answer comes back. You see the result in your chat, as if the AI had done it itself. If the API needs a key, the helper adds it quietly so your AI never sees it.

Click to zoom

What you can do with it

• Look up live data from any API that has an OpenAPI or Swagger file
• Send information to an API, like creating or updating a record
• Filter which API actions your AI is allowed to use
• Keep your API key hidden from the AI assistant
• Point at a local spec file or a spec hosted online
• Add custom headers to every request when a service needs them

Try asking your AI

• “What is the weather forecast for Paris tomorrow?”
• “Find the current price of item 1234 in the store API”
• “Create a new support ticket titled Printer is broken”
• “List the last five orders from the shop API”

What it gives back to you

You get the API's answer back in your chat, usually as text, a small table, or a list of items. If the action changed something on the other side, the helper tells you what it did. Sometimes it is just a short confirmation, like a new ticket number. The exact shape depends on what the API sends back.

Before you start

Who it's for

People who want their AI to work with an outside service that already publishes API documentation, without writing any code.