By PAYATE
You try to copy a sentence from a PDF and nothing happens - or you get strange characters. The PDF is a scan: a picture of paper, with no real text inside. OCR (optical character recognition) reads the picture and turns it back into text you can copy, search and edit.
Table of contents
How to extract text with OCR
- Open OCR PDF. It accepts scanned PDFs, photos (JPG, PNG) and screenshots.
- Choose the document's language - English, Arabic, Arabic + English, French, Spanish and more.
- For long PDFs, choose the pages, for example
1-5. - Click Recognize text. The first time, the engine downloads (about 10 MB).
- Copy the text, or download it as TXT or Word.
Recognition runs on your own device, so private documents such as contracts and IDs are not uploaded.
How to tell if a PDF is scanned
- You cannot select words with the mouse.
- Search (Ctrl+F) finds nothing, even for words you can see.
- Zooming in shows blurry, pixelated letters.
If you can select the text, you do not need OCR - PDF to Text or PDF to Word is faster and 100% accurate.
Getting accurate results
- Resolution: 300 DPI scans or sharp phone photos give the best results.
- Straight pages: rotate sideways pages first with Rotate PDF.
- Contrast: for faded copies, rescan with Scan to PDF in Document mode.
- Language: the right language setting matters most - Arabic text needs the Arabic option.
OCR for Arabic documents
Arabic OCR is harder because letters join and change shape. Choose "Arabic" or "Arabic + English" and use a clear scan.
Frequently asked questions
Is OCR free?
Yes, our OCR is free with no daily page limit.
Why are some words wrong?
OCR depends on scan quality. Blurry, tilted or low-contrast pages cause mistakes; rescan or straighten the page.
Can OCR read handwriting?
Printed text works best. Neat handwriting may be partly recognised, but results vary.
Are my scans uploaded?
No, the OCR engine runs inside your browser.
Try OCR: Extract text now →
