FastAPI

Backend for close to a dozen services at this point (FanDesk, our CMS, a couple of internal tools) - all Python underneath. The auto-generated OpenAPI docs alone save a stupid amount of frontend/backend back-and-forth once more than one person touches an API, and Pydantic catching a bad payload before it reaches business logic has probably saved more debugging time than any other single framework choice we've made. Only real friction: mixing sync and async bites you occasionally if a dependency isn't actually async under the hood, and it shows up as a mysteriously stalled event loop instead of a clear error.

We switched to FastAPI for our core services, and the performance gain was immediate. The automatic interactive documentation (Swagger UI) is a lifesaver for team collaboration. Type hinting with Pydantic makes the code much cleaner and reduces bugs significantly compared to Flask. If you're building a data-heavy application where speed and developer experience matter, this is the gold standard for Python.

I like FastAPI, because the documentation is not a nerd explaining to nerds, but an expert explaining to normal people. I am still going through the documentation, so I can not say much more :-)

fast, easy to code and intuitive

Really awesome :)

It is exceptionally clean, does wonders to object serialization through Pydantic, is exceptionally performant, very easy to configure a production-ready version. I was completely swayed by documentation from the first time I saw it. Great work all around.