By definition, OCR means optical character recognition. It depends on the contents of the PDF what kind of extraction methodology can work. Often some available PDFs are just scans of printed documents or handwritten notes. If machine readable text is available your approach is great.
mindblowing tool. i've attempted to create a react app to plan meals. each meal consists of several recipes, each recipe of several ingredients. user can edit ingredients with their nutrition values and build recipes from multiple meals etc. then added a navigation bar with multiple pages and a sqlite database. after adding the sqlite database I ran out of free tokens and there were some errors preventing the app from running. but until this point there was a functional app to edit ingredients, recipes and mealplans without data persistence that was created in ~15minutes. pretty cool! i'd love to play more with it, maybe tomorrow if I get some free tokens again.