PDF OCR

OCR (optical character recognition) analyzes the pixels of a scanned page and recognizes which shapes are which letters, turning a picture of text back into actual, selectable text. A scanned PDF — even one that looks completely normal on screen — is really just an image as far as your computer is concerned: you can't select it, search it with Ctrl+F, or copy a sentence out of it, until OCR has run on it.

• Digitizing paper archives — old contracts, receipts, or books that only exist as scans.
• Making a scanned form searchable — finding one clause in a 50-page scanned agreement without reading every page.
• Preparing a scan for PDF to Word — OCR has to run first, or there's no real text to convert.

Accuracy is very high on clean, well-lit scans of typed text in a supported language — often close to perfect. It drops noticeably on blurry or skewed scans, handwriting (which OCR generally can't read reliably at all), unusual fonts, and documents scanned at low resolution. Selecting the correct document language before running OCR matters more than most people expect — recognition is language-specific, and picking the wrong one will produce garbled results even on a perfectly clean scan.

• Scan at a reasonable resolution (200–300 DPI) — too low loses detail OCR needs, too high just adds file size without helping accuracy.
• Straighten skewed scans before running OCR where possible — a tilted page reduces recognition accuracy.
• Select the correct language — this is the single most common cause of poor OCR results on an otherwise clean scan.