1. Allow pasting a picture in addition to OCR from file.
2. Allow OCR of a fragment of the screen (combine partial screen capture with OCR).
Option 1 combined with partial screenshot (even the one built-in in MacOS or Windows) would be a workaround for Option 2.