Text recognition (OCR)
A scan is just a picture — recognize its text to search, copy and reuse it with other tools, right on your device.
Add files to start
Drag and drop files here, or
Files are processed right on your device.
Recognition
PDF pages that already have a text layer (not scans) are kept as they are.
Runs on your device with Tesseract (WebAssembly). Each page takes a few seconds; slower devices take longer.
With text available, you can continue with: Compare PDF · Redact PDF
How to use
- Add a scanned PDF or photos of documents (each photo becomes a page).
- Choose the language on the document (English, Spanish, French, Indonesian, Vietnamese); tick “Also contains English text” for mixed documents.
- Press “Recognize text”. The first time, the recognizer is downloaded from this site (a few MB).
- Download the PDF with its new text layer (original pages unchanged) or copy/download the text as .txt.
Current limits
- Works best with clear printed text, straight images and about 300 dpi. Handwriting, blurry or strongly skewed images and very small text give weaker results — low-confidence pages are flagged.
- The text layer is invisible and placed at each word’s position; page images do not change. Tables and multi-column layouts can make the text order differ from reading order.
- It runs on your device, so speed depends on it: usually a few seconds per page; older phones may be slower.
Frequently asked questions
Are my files uploaded for recognition?
No. The Tesseract recognizer runs in your browser (WebAssembly). The site only downloads the recognizer and language data; your files and page images never leave your device.
Does the PDF lose quality?
No. Original pages are kept and scanned images are not recompressed; only an invisible text layer is added for search and copy.
What if the text is wrong?
Proofread it, especially pages flagged with low confidence. Clear, straight, well-lit images give better results — you can flatten phone photos in Prepare documents first.