PDF
Runs an on-device AI model over every page of your PDF to recognize text, even on scanned documents with no selectable text layer, and exports the result as a Markdown file.
How to use OCR PDF
- Add a scanned PDF.
- Pick the document's language.
- Tap Run — the model downloads once, then reads each page fully on-device.
- Download the .md file with the recognized text.
FAQ
Does this upload my file anywhere?
No — the OCR model runs entirely in your browser. It downloads once (a few megabytes) and then works fully offline; your PDF never leaves your device.
What languages are supported?
English, Spanish, French, German, Portuguese, Hindi, Arabic, and Simplified Chinese. Pick the one that matches your document before running.
How accurate is it?
Accuracy depends heavily on scan quality — clean, high-contrast, upright scans do best. Handwriting, skewed pages, and low-resolution scans will have more errors.