Anonymize a PDF without uploading it
Drop your PDF. ONYRI Sanitize finds the names, ID numbers, bank details and contact data inside it, masks them, and gives you back the same PDF with the same layout. Everything happens in your browser. No server ever sees the file.
Runs 100% in your browser. Your file is never uploaded.
What ONYRI finds in a PDF
- Names of people, in signatures, headers and body text
- Email addresses and phone numbers
- Postal addresses and postcodes
- Official ID numbers: passport, national ID, social security
- Bank details: IBAN, BIC, card numbers
- Company identifiers: VAT number, SIREN, SIRET, EIN
- Dates such as birth dates and signing dates
- With Pro: API keys and technical secrets pasted into reports
How to anonymize a PDF in three steps
- 1
Drop the PDF
Drag the file onto the page or pick it from your disk. ONYRI reads the text layer of each page right in your browser. A scanned page with no text layer is read by on-device OCR instead. Nothing is sent anywhere.
- 2
Review the preview
Every detected value is masked on the page, in place. Untick anything you want to keep. Type a word under “Also mask” to hide every occurrence the engine missed.
- 3
Download the masked PDF
You get a new PDF with the same pages and the same layout. The export is flattened: each page becomes an image, so no hidden text layer sits under the black bars.
Why a PDF should be anonymized on your own device
Most online PDF tools start with an upload. Your contract, your report or your bank statement lands on a server you do not control, before any masking happens. That upload is the very risk you were trying to remove.
ONYRI works the other way round. The PDF is opened, read and masked by code that runs in your browser tab. You can check it yourself: open the developer tools, watch the Network tab, and run a full cycle. No request carries your file.
A classic mistake with PDFs is drawing a black box over the text in a PDF editor. The text is still there underneath, and anyone can copy it. ONYRI exports a flattened file, so what is masked is really gone from the copy you share.
What you should know before you share the file
- Detection relies on patterns and rules. It is not exhaustive. Always read the preview before you download, and add by hand what the engine missed.
- Scanned pages with no text layer are read by on-device OCR (English, French, German), page by page. This is slower, and accuracy depends on the scan quality. Handwriting is not read. Check these pages with extra care in the preview.
- The exported PDF is made of page images. The text can no longer be selected or searched. Keep your original file if you still need it.
- Black-marker mode removes the data for good. Token mode is pseudonymization: under the GDPR, pseudonymized data is still personal data. Anonymizing a file reduces exposure; it does not by itself make you compliant.
Anonymizing PDFs: frequent questions
Is my PDF uploaded to a server?
No. The file is read and masked by code running in your browser. ONYRI's servers only handle your account, your plan and your saved detection settings. They never receive the document or its content.
Can someone remove the black bars and read the text?
Not in the exported file. The export is flattened: each page is rebuilt as an image with the bars painted in. There is no text layer under them to copy or to reveal.
Does the layout stay the same?
Yes. Pages keep their size, fonts, tables and logos. Only the detected values are covered, exactly where they appear.
How many pages can I process?
The Free plan covers 3 documents a day, up to 10 pages each, in black-marker mode. Pro removes the document and page limits and adds token mode and technical-secret detectors.
What is the difference between the two masking modes?
Black marker covers each value with a solid bar. It cannot be undone. Token mode, on Pro, replaces each value with a label like [NAME1], so the document stays readable. The same person always gets the same label.
Does it work with a scanned PDF?
Yes. When a page has no text layer, ONYRI reads it with the same on-device OCR used for images, in English, French and German. Expect a few seconds per page. OCR can misread a blurry word and does not read handwriting, so check the preview before you download.
Your document, made confidential in seconds.
Drop a file, check the preview, download the masked copy. The file never leaves your device.