Does my scan get uploaded for recognition?

No. The recognition engine and the language models are downloaded to your browser, and the pages are read on your own device.

Why is the first run slower?

The engine and language model download once (a few MB) and are then cached, so later documents start immediately.

How good is the Arabic recognition?

It handles clear, printed Arabic well. Handwriting, low-resolution scans and heavy decoration will always be harder for any OCR engine.

OCR — scanned PDF to text

Turn a scanned or photographed PDF into editable text. Arabic and English are both supported, and recognition runs entirely on your device — the pages are never uploaded.

or drop files here

How it works

1
Select a scanned PDF.
2
Choose the document's language — Arabic, English, or both.
3
Click OCR and download the recognised text, or copy it straight from the page.

Frequently asked questions

100% private — files never leave your device