I dont know how difficult this would be to implement, but just adding onto the feature requests for future undertaking!
USE CASE :
I find myself creating notes with images more than writing texts. For example, i have my textbooks in a pdf format opened in browser while i make notes. So whenever i find some content which is worthy of being stored in my notes, i screenshot that paet of the pdf and paste it in anytype. Like this, i have almost all the notes with JUST images containing texts which is a way more faster way to store notes. If i would write all of it in texts, that would take a lot of time to build up my college notes. Having my notes implemented in form of images mostly, I cant search for useful information through universal search, almost none of them. I need to remember where the info is stored, navigate there manually and then read it. OCR would be a game changer for this!