Paperless 3 adds significant improvements to the OCR functionality versus Paperless 2. Paperless 3 uses a new and very customized  OCR engine that looks at glyphs using separate image libraries based on the font.   Specifically, Paperless will look for certain keywords that match our fields, even if they are slightly misspelled. It automatically stores and compiles a list of those words so that on future scans Paperless will rescan and if it senses a problem, will rescan. Our testing in QA has shown, in some cases,  pretty dramatic improvements in the amount of accuracy we’ve achieved since version 2.