> can be hard to extract text from them with high fidelity. Or impossible with scans. Usually I take a look at the filesize as a guestimate. Anyone know better

> can be hard to extract text from them with high fidelity. Or impossible with scans. Usually I take a look at the filesize as a guestimate. Anyone know better FOSS OCR tools that run from cmdline?