I noticed in the release notes something about using iOS’ built-in text recognition abilities in DTTG now. Does that mean DTTG no longer uses the ABBYY engine and just uses the iOS one? If so, that’s great – I’ve found iOS’ text recognition to be far and above any other OCR engine out there.
Am I understanding this all correctly? Thanks!
DEVONthink To Go never used he ABBYY engine, only DEVONthink for Mac does. On iOS, we’re using iOS’ own Vision framework for the languages it supports and our own flavor of the Tesseract engine for all other languages and scripts.
That makes sense – thanks for the explanation! DTTG’s OCR is second-to-none in my book. Any chance the Mac version could use iOS’ vision framework? 
I’d like to point out that Vision still supports fewer languages than Abbyy‘s OCR. And is, in my experience, sometimes not rising to the challenge of assembling words/lines in the correct sequence.
This is one of the weaknesses of Vision, it has no always a good understanding of especially multi-column documents.
That’s all good to know – I guess my use case of OCR relies more on just being able to search for a few keywords to find a PDF, handwritten note, etc. iOS Vision seems to be great for that, but I can empathize with needing the more structured output ABBYY provides.