OCR a PDF — make scanned pages searchable
Scanned contracts and photographed pages are pictures: you can’t search them or copy a sentence. OCR reads the words and adds a hidden text layer under the image.
Drop a PDF file here
Free · no signup · no watermark · files stay on this device · how local processing works
How to use OCR PDF
- Open the scanned PDFPages that already contain text can be skipped automatically.
- Choose accuracyBalanced 200 dpi works for most office scans; Accurate for small print.
- Run OCRDownload a PDF that looks the same but is searchable.
Runs locally — and what that means for speed
The Tesseract OCR engine is compiled to WebAssembly and runs in your tab. The first run downloads the engine and the English model (about 3 MB in total), after which they are cached. Expect roughly 2–6 seconds per page on a laptop and longer on phones. In exchange, your scans never leave your device.
What the output looks like
The original page image is kept untouched. Recognized words are placed exactly over their position as invisible text, so the PDF looks identical but Ctrl+F finds words, you can select and copy text, and search engines or document systems can index it.
Getting good results
OCR accuracy depends on scan quality. Straight, high-contrast scans at 200–300 dpi give the best results; skewed phone photos and decorative fonts give the worst. After OCR, PDF to text pulls the recognized text into a .txt file, and Compare PDFs can then diff two scanned versions.
Need more than one step? The Workbench runs OCR PDF together with other operations on the same file in one go — for example merge → compress → protect.
Questions about OCR PDF
Which languages are supported?
English in this version. Other Latin-script text is often partly recognized; more language models are planned.
Why is it slower than online OCR services?
Those services run on servers after you upload your file. Here your own processor does the work, so nothing is uploaded.
Will OCR change how my PDF looks?
No. The text layer is invisible; the page images are not modified.
Can I OCR a PDF that already has some text?
Yes. With “Skip pages that already have a text layer” on, only image-only pages are processed.
How long does OCR take?
Roughly a few seconds per page on a recent laptop at the balanced setting, longer on phones and at 300 dpi. The progress line shows the page being recognized, and pages that already have a text layer can be skipped to save time.