@hhursev I ran recipe-scrapers against a Pinch of Yum cookie recipe and a BBC Good Food pancake page. Both returned clean structured data — title, total time, yields, full ingredient list, step-by-step instructions — in a few lines of Python.
The design choice that stands out: each of the 376 supported sites has its own scraper class, but the caller never picks one. scrape_html resolves the right parser from the URL, so the same three-line script works across AllRecipes, BBC Good Food, Jamie Oliver, and hundreds more. That URL-in → structured-recipe-out contract is what makes the library composable for meal planners, grocery list builders, and nutrition trackers.
The gap is that a repo visitor who just wants to try it still needs Python and pip. I set up an online entry point where anyone can paste a recipe URL and get the structured output directly:
Paste a recipe URL and try it here — it runs the same extraction and returns the structured fields without any local setup.
Every session that comes through the link leaves a usage record on the platform, so you can see which sites people actually test and where the parser trips. That signal feeds straight back into scraper coverage decisions.
Feel free to close if this isn't relevant.
shesonglin@tinkerland.ai
@hhursev I ran recipe-scrapers against a Pinch of Yum cookie recipe and a BBC Good Food pancake page. Both returned clean structured data — title, total time, yields, full ingredient list, step-by-step instructions — in a few lines of Python.
The design choice that stands out: each of the 376 supported sites has its own scraper class, but the caller never picks one.
scrape_htmlresolves the right parser from the URL, so the same three-line script works across AllRecipes, BBC Good Food, Jamie Oliver, and hundreds more. That URL-in → structured-recipe-out contract is what makes the library composable for meal planners, grocery list builders, and nutrition trackers.The gap is that a repo visitor who just wants to try it still needs Python and pip. I set up an online entry point where anyone can paste a recipe URL and get the structured output directly:
Paste a recipe URL and try it here — it runs the same extraction and returns the structured fields without any local setup.
Every session that comes through the link leaves a usage record on the platform, so you can see which sites people actually test and where the parser trips. That signal feeds straight back into scraper coverage decisions.
Feel free to close if this isn't relevant.
shesonglin@tinkerland.ai