I think I've got the wrong end of the stick here, but I user a product called Paperport for scanning random bits of paper. It's got OCR built in, but not accessible to the user: when you scan something, it does OCR in the background to create a searchable index of your documents. So, if I search for "rubber duck", it brings up all the documents containing that term. When I click on the document, I see the original version as scanned, not an OCR'd version. It's one of the few applications that has moved computers with me.
I've got a very fast sheet-fed double-sided scanner, which makes a huge difference, too: receipts, magazines, photos and pretty much everything else get dumped through it.