What's this about?
A small tool for extracting text from documents you cannot select or copy from — scanned PDFs, locked viewers, images — so the text can be handed to a model.
What exactly is OCR?
Optical character recognition turns a picture of text back into text. It has existed for decades, but it has become newly important because models need actual text as input, so any document trapped as an image is invisible to your AI tooling until something converts it.
What's the community saying?
HN 139 points, 36 comments
Mostly comparisons with existing OCR options and discussion of accuracy on difficult layouts like tables and multi-column academic papers.