My understanding is that at the moment there is no Codechef API, you have to parse web pages

I am afraid that your idea is much similar with https://clist.by, which also provides APIs to fetch user statistics. If you only want to check contest datas of leetcode and codechef, maybe my website, https://chinesedfan.github.io/clist-ex/, can give some help.

I ran into a very similar issue while building APIs for my own puzzle-solving platform, especially when trying to keep response times low for screenshot processing and solver requests. One thing that helped a lot was separating the scraping/discovery layer from the actual FastAPI endpoints instead of directly coupling parsing logic with request handlers. That made retries, caching, and rate-limit handling much cleaner.

For example, I ended up structuring endpoints like this:

from fastapi import FastAPI app = FastAPI() @app.get("/solve") async def solve_board(board_data: str): result = solver_engine(board_data) return {"solution": result}

Then a separate async service handled: • request retries • endpoint rotation • parsing • fallback responses • cache refresh

For sites like LeetCode/CodeChef, sometimes inspecting: • XHR/fetch requests in DevTools • GraphQL calls • contest APIs • ranking endpoints works better than looking at page HTML itself. I also learned pretty quickly that keeping API responses lightweight matters a LOT once traffic scales. My Block Blast Solver project started getting much faster after introducing async request batching and cleaner endpoint isolation.

Especially if you're planning real-time solving or competitive stats fetching, FastAPI + async architecture is honestly a very solid choice.