I import a lot of older books and text images from classes, but in order to use goodnotes AI, the words in the imported PDF need to be actual text.

MacOS does it well in preview, but I can't export it as embedded text from the image scan, so I'll have to resort to janky online tools that take forever to read the images then ultimately import it again to goodnotes.

I've seen some occassional messages from the AI sidebar asking if it can read it, but it never does and the message is sporatic anyway. How can goodnotes read sloppy handwriting but not text from an image or book??