It would be great to have OCR mode in LO Writer. I could image two ways of using it: 1. recognition mode - something like importing pdfs or images, declaring size for page (optional); there would be two columns - for picture and for text document. the picture area could be sent to unpaper or manually could have drawn boxes for ocr. then the boxes areas would be sent to an ocr engine and the result would be put into text windows (into frames?) 2. layered mode - a text document, you import graphics manually, and select graphical area for ocr. the result returned from an ocr engine would be in new layer (not sure if LO Writer supports layers somehow.)
Sounds like a job for an extension. Otherwise we would have to pull in Tesseract as a dependency for all LibreOffice users. On the other hand, there are free software solutions that have a graphical interface and support exporting to ODT: https://github.com/zdenop/lector Asking UX for their opinion. I would say WONTFIX all thing considered.
We're replacing our use of the 'ux-advise' component with a keyword: Component -> LibreOffice Add Keyword: needsUXEval [NinjaEdit]
Writer's scope is surely not to be the swiss-army knife. And the interest for this idea was low in the last year, so I'll close it as WONTFIX.
I'd like to see this as well, but it is out of scope for core of LibreOffice. Fertile grounds for an Extension (hopefully FOSS and cross platform). Concur with the WONTFIX
*** Bug 151957 has been marked as a duplicate of this bug. ***
OP requested Writer but function would be all pdf import filters (Draw, Impress, Writer). But still out of scope for core.