Anonymize a scanned PDF, with OCR in your browser
Recognize a scan's text with local OCR and replace personal data with reversible placeholders. Neither the image nor the text leaves your device.
- No sign-up
- Free, no watermark
- Nothing leaves your device
Sent to servers: 0 BNothing left your device
About this tool
A scan is a photo of each page: there is no text to anonymize until it is recognized. Here OCR runs in your browser, with a model of about 3 MB downloaded once from this same site.
Then the recognized text goes through the same anonymization as a digital PDF. With skewed or low-quality scans OCR makes more mistakes, and a misread value may not be detected: review the result more carefully than with a digital PDF.
How it works
- 01
Drop the scanned PDF.
- 02
Accept text recognition (OCR) when the tool offers it.
- 03
Review the text and each highlighted item.
- 04
Copy or download the result.
Questions
Do I get a blacked-out PDF?
No, on purpose. The result is text (Markdown) with placeholders such as [NOMBRE_1] instead of the data. We do not do destructive redaction: nothing is destroyed, and with the map you can put the real data back into the AI's answer.
Does it catch every piece of personal data?
Not always. It works by rules: ID and tax numbers, emails, phones, dates, addresses and names when labelled. An oddly formatted value can slip through: review the result before sharing.
Is OCR done on a server?
No. The engine and the language model are downloaded once from this site and run in your browser; afterwards they work offline.