I work with an OJS install where some digitized back issues of journals have the entire run as PDF, but no one ran OCR on them. I noticed I can use SSH to download the files for them to my desktop, run OCR on them, then use SSH to replace the PDF on the server with the OCRed one. If I do this will anything go wrong as far as the software and people being able to download and read the PDFs goes?
(I realize that this might not be great for provenance, and that if I down sample images, I can’t go back.)