A scan can look readable yet contain no searchable text. NAPS2 can add recognized text to these PDF pages. The page image alone does not prove that names, numbers or reading order were recognized correctly.
Keep the original and inspect pages
Work on a copy and preserve the original PDF. Import the copy into NAPS2. Check page count, order and orientation. For cropped lines, heavy shadows or blurred characters, a better scan often helps more than another OCR run. This guide is for scanned documents, not for preserving interactive form functions or digital signatures.
Configure OCR and save as PDF
- Open the OCR button in NAPS2; on Mac the documentation places it under Tools, OCR.
- Download the required recognition language on first use. Choose the document language, not automatically the interface language.
- Enable the option to make PDFs searchable using OCR. For multilingual documents, download the language packs and select multiple languages.
- Save a new PDF with a distinguishable name. Do not overwrite the original; wait until recognition and saving finish.
NAPS2 normally runs OCR during saving. Best mode may improve results compared with Fast but takes longer. White-balance and noise options can help weak scans; still inspect the result.
Read back the saved result
Close the saved PDF and open that exact file in a PDF viewer. Search for a distinctive word on an early and a late page. Select a paragraph, copy it into a local text editor and compare it with the visible scan.
Pay particular attention to names, accented characters, dates, decimal separators and confusions such as O/0 and I/1. Check reading order in columns and tables. Important amounts or identifiers need complete verification; samples do not confirm every page. A search match proves only that match.
Existing text and limitations
For imported PDFs, NAPS2 documents that pages already containing text are left alone, while pages without text receive OCR. Reimporting therefore does not automatically correct an existing faulty text layer. Determine which pages are actually image-only instead of repeatedly processing the same file.
OCR does not translate or automatically create an accessible PDF or a correctly formatted text file. NAPS2 does not export OCR results directly as text; its documented route is saving PDF and copying text in a viewer. Retain the original and document unresolved recognition errors before sharing.