FastAPI

The FastAPI framework has become very popular in recent years due to its speed comparable to APIs written in Go or NodeJS. It is most often run on asynchronous servers like Starlette or Pydantic.

APIs built with FastAPI are written quickly due to concise code, minimal duplication, and good documentation. FastAPI is fully compatible with the OpenAPI standard.

Building a simple API is limited to 9 lines of code.

from typing import Optional

from fastapi import FastAPI

app = FastAPI()


@app.get("/")
async def read_root():
    return {"Hello": "World"}


@app.get("/items/{item_id}")
async def read_item(item_id: int, q: Optional[str] = None):
    return {"item_id": item_id, "q": q}

FastAPI has a built-in documentation module

FastAPI has a built-in module that automatically generates documentation in both Swagger and ReDoc formats. By making changes to data models using BaseModel from Pydantic, we get updated and described documentation under url: http://127.0.0.1:8000/docs.

FastAPI - speed, security and a large community

Using modules such as oauth2, http, PyJWT we can get application security such as OAuth2, or JWT token security. In addition, using appropriate middleware we can impose other security standards, these include CORS (Cross Origin Resource Sharing).

The popularity of open-source solutions is measured by the number of contributors to a given application. In the case of FastAPI, there are 284 active contributors, as well as 50,000 people actively tracking changes to the framework. By comparison, the once most popular Django has a community of 61,000 people actively following changes, as well as 2156 people actively contributing to changes in the framework.

In addition, in independent tests conducted by https://www.techempower.com/benchmarks/#section=data-r20&hw=ph&test=composite&l=dbf0n3-7b&a=2&f=zik0zj-qmx0qn-zhwum7-zik0zi-z8kflr-yyku7z-zik0zj-zijunz-zik0zj-zik0zj-zik0zj-1kv, it gets one of the highest places among the frameworks around which the extensive ecosystem of solutions is based. There are faster solutions, but certainly not with such a community of developers and available plug-ins.