Data and backend

FastAPI

Typed Python web framework with generated OpenAPI.

made by
FastAPI

Magic Ship is one shop in Vancouver, BC, working remotely with clients worldwide. We are not a partner, reseller, or certified vendor of FastAPI - we just build with this.

What FastAPI is

FastAPI is an async Python web framework built on Starlette and Pydantic. Request and response models are Pydantic classes, so validation and the OpenAPI schema both fall out of the type annotations. Streaming responses and background tasks are built in.

How we use it

It is our default for Python services in an AI product: the endpoint in front of a model, the ingestion API, the handler a queue calls. Streaming responses carry token-by-token output to the browser over server-sent events, and the generated OpenAPI schema is what we generate a typed TypeScript client from, so the two sides cannot drift. The same Pydantic models get reused as tool schemas, which keeps the API and the agent agreeing on shapes.

Where it is the wrong choice

It punishes CPU-bound work: one synchronous handler that blocks the event loop quietly serializes every request behind it, and the fix is a worker process, not a framework. For a team that is entirely TypeScript, adding Python to the deploy story usually costs more than it returns.

Building something on FastAPI?

Send the problem rather than a job spec. You get an answer on scope, on fit, and on whetherFastAPI is even the right call for it.

Start a project

FastAPI and FastAPI are trademarks of their respective owners, used here to say what we work with.