A scanned contract, a page from a book, or a photographed certificate saved as a PDF looks like a regular document, but you can't select, copy, or search for the text in it — because it isn't text, it's an image. To fix that, you need text recognition, or OCR.
Why scans don't have text
A PDF can store a page in two fundamentally different ways: as actual text (characters with a font and coordinates) or as an image — a photo or scan of the page packed into a PDF wrapper. In the second case, it's just a picture as far as a computer is concerned, even though a human reads it as ordinary text. Files like that are exactly what OCR is for.
How to recognize text in a PDF online
The OCR for PDF tool renders each page as an image, then recognizes the text using the Tesseract OCR engine in Russian and English. Upload the file, choose the output format, and download the result — processing happens entirely on the server.
Two output formats
- TXT — the plain extracted text, with no markup or formatting. Good when all you need is the text itself: to copy it, find a specific passage, or process it further.
- PDF with a text layer — the same document as the input, but with an invisible layer of recognized text placed over the page images. It looks like the original, but the text can be selected, copied, and searched with Ctrl+F — handy when preserving the document's appearance matters.
What determines recognition quality
OCR accuracy depends directly on the scan's quality: sharp, evenly scanned printed text gets recognized nearly flawlessly. Blurry, rotated, or photographed-at-an-angle pages produce more errors. Handwritten text is recognized noticeably worse than print — Tesseract OCR is primarily built for printed text.
Bottom line
OCR turns a "mute" scan into a document with real text — ready for copying, searching, or further processing. A TXT result is convenient for working with the text itself, while a PDF with a text layer preserves the original's appearance while adding the ability to search and copy.