You open a scanned PDF, try to search for a date, and nothing comes up. That's because the file is a picture, not text. OCR — Optical Character Recognition — is the trick that converts those pictures of words into real, searchable text.
How OCR sees a page
OCR software analyses each shape on the page and matches it against patterns for letters and numbers. A capital 'A' has predictable angles; a '7' has its own profile. Once recognised, the software stitches characters into words and lines. Modern OCR uses AI to handle tricky bits like tilted scans, faint print and old fonts that would have stumped earlier tools.
What changes after OCR
Before OCR: you can see the words but not select them. After: you can copy text, search the file, and convert PDF to Word for editing. The page still looks identical — OCR adds an invisible text layer behind the image. You get both: the original look of the scan, plus the ability to interact with the text.
When you'll need it
Any time you're working with a scanned PDF and need to do more than just look at it. Searching old receipts, pulling figures out of a report, redacting a personal detail — all of these need OCR first. If your PDF was made from a Word document, it already has selectable text and doesn't need OCR.
FAQ
Is OCR always accurate?
Modern OCR is around 99% accurate on clean typed text. Accuracy drops on handwriting, smudged paper, unusual fonts and busy backgrounds. Always proofread important sections before relying on the result.
Will OCR change how the PDF looks?
No. OCR adds an invisible text layer behind the page image. Visually the PDF looks identical to before, but now you can select and search the text.
Does OCR work on handwriting?
Sort of. Clear printed handwriting can be recognised by modern tools, but cursive script and rushed notes are still hit-and-miss. For typed scans, expect excellent results.
Got a scan you need to search or edit? Run it through Flint and you'll have selectable text in moments.