Thank you for your comment. Currently, there are three methods of adding a book: searching (which uses the Google Books API under the hood), manual addition (which nobody wants to do — I completely understand — but is the only option for old books that cannot be found online; I have a lot of these) and scanning the ISBN (which also uses the Google Books API). I initially thought about adding the method you suggested, where you provide a picture and then extract the information, but I decided it was not the right time to implement that. So, I left it for later, when there is any need or feedback regarding it from users. From your comment, I understand that it's something that users might actually consider using.

Yeah if there is an ISBN listed, that will be fastest. But for older books that don't have an ISBN, if there is a title on the cover or spine, OpenAI is really good at extracting that, regardless of the font or image quality. It's like a super OCR. From the title and author you can probably stub in the rest of the data, and the user can correct it if necessary.