Two PDFs can look identical on screen and be completely different inside. One contains text: letters the computer knows are letters, which you can select, search, copy and have read aloud. The other contains a photograph of a page, and to the computer it is just a grid of coloured dots. The difference decides whether you can search the document, copy from it, fill it in, shrink it, or redact it properly.
How to tell in five seconds
- Try to select a word. Click and drag across a line. If the text highlights word by word, the PDF has text. If you get a box around the whole page, or nothing, it is a scan.
- Search for a word you can see. Press Ctrl+F (Cmd+F on a Mac) and type a word on the page. If it is not found, there is no text layer.
- Zoom right in. Text in a real PDF stays perfectly sharp at any zoom. A scan becomes blurry or blocky.
What OCR does
Optical character recognition looks at a picture of text and works out which letters it shows. Good OCR adds an invisible layer of recognised text behind the image, so the page looks exactly like the scan but can be searched and copied. It is called a "searchable PDF".
OCR is never perfect. It confuses similar shapes, such as 0 and O, 1 and l, and rn and m, and struggles with handwriting, low-resolution scans, stamps over text and unusual fonts. Accuracy on a clean, straight scan at 300 DPI is very high; on a dim phone photo it falls quickly.
Free ways to make a scan searchable
Google Drive. Upload the PDF, right-click it and choose Open with, Google Docs. Docs runs OCR and gives you the recognised text as an editable document. The layout is lost, but the words are there to copy. It supports many languages, including Devanagari scripts.
Phone scanner apps, such as the scanner in Google Drive, often add recognised text automatically when they save a PDF.
The tools on this site do not include OCR yet. PDF to text extracts text from PDFs that already contain it, and tells you when a PDF has none.
When it matters
Searching long documents. Finding one clause in a 200-page scanned contract means reading all of it, unless it has been OCR'd.
Accessibility. Screen readers used by blind and partially sighted people read the text layer. A scan without one is silent.
Redaction. Searching for every occurrence of a name only works on text. On a scan, mark each one by hand; Redact PDF tells you when a file has no text layer.
File size. A text PDF of ten pages may be 100 KB; the same pages scanned can be 10 MB. If you have the original, export it to PDF rather than printing and scanning it. If you only have the scan, Compress PDF makes it manageable.
Avoid making scans you do not need
The most common source of unsearchable PDFs is the print-sign-scan routine: a perfectly good digital document is printed, signed with a pen and scanned back in. Signing digitally with Sign PDF keeps the text layer intact, keeps the file small, and skips the printer.