Anonymize PDFs on your computer

Automatic detection of personal data (IDs, names, phone numbers, IBANs, signatures, faces and 100+ patterns) with OCR for scanned documents. Your files never leave your device. Free includes one PDF per day; the 7-Day Pass and Pro are unlimited.

100% private: everything runs on your computer, your files never leave your device.

  1. Upload your PDF or image - nothing is uploaded to any server
  2. Review the detected personal data
  3. Download the anonymized PDF

What Anonymize PDF does

Anonymize PDF finds personal data in a document and removes it for good. It reads the text of every page, and for scanned pages it first recognises the text locally, then looks for identity numbers, names, phone numbers, email addresses, IBANs, card numbers, postal addresses, health data, company identifiers, handwritten signatures, QR and bar codes and faces. You review every detection, decide how each category is redacted, and download a PDF in which the redacted areas are physically gone, not covered.

It is built for the situations where sharing a document means sharing someone else's data: a court file sent to a third party, a medical report used for a second opinion, a contract shared with a consultant, a CV or a bank statement attached to an application. The detection engine, the text recognition and the redaction all run in your browser.

How to anonymize a PDF

  1. Add the PDF or the image. Before starting you can choose the Detection precision: Normal, the recommended default, or Maximum, which loads an additional name recognition model of about 130 MB from OléPDF's own domain and takes longer.
  2. Wait while the pages are analysed. Scanned pages go through local text recognition first, and the progress screen tells you which page is being read.
  3. Review the detections. They are grouped by category, each one is highlighted on the page, and you can deselect anything that should stay, search the document to add more, or draw areas by hand with the blackout, crop, asterisks, mask or custom text tools.
  4. Choose the Redaction style per category: Blackout, Gray blackout, Clip, Asterisks, Partial, which leaves the last few characters visible, or Custom text, which prints a replacement such as [REDACTED].
  5. Press Generate Anonymized PDF and confirm. The file downloads with the suffix _redacted and a timestamp. Always check the result before sharing it.

How the redaction works and what its limits are

Detection combines more than a hundred pattern rules for identifiers from Spain, the European Union, the United States, the United Kingdom, France, Germany, Italy, Portugal, the Netherlands, Mexico and Brazil with structural rules for names in lists and tables, a local face detector and a code reader for QR and bar codes. Pages that contain at least one redaction are rendered as images with the boxes painted in, which is what makes the removal irreversible; pages without redactions are copied unchanged.

Privacy: nothing leaves your device

The document, its text, the page images and the detections stay in your browser tab. Everything the tool needs, the pattern rules, the text recognition engine, the face detector, the code reader and the optional name recognition model, is downloaded from OléPDF's own domain and runs on your computer. No page, no text and no detection is sent to a server.

That is the whole point of the tool: you are anonymizing a document precisely because its content is sensitive, so it would make no sense to upload it in order to do so.

Typical uses

Frequently asked questions

Is Anonymize PDF free?

The free plan includes one anonymized document per day for this tool, with the full detection and review workflow. The 7-Day Pass and Pro make it unlimited and add batch processing; the audit report requires Pro.

Is the redaction really irreversible?

Yes. Pages with redactions are rebuilt as images with the areas painted over, so the original text and pixels are no longer in the file. This is different from placing a black box on top of the text, which can be removed.

What is the difference between Normal and Maximum precision?

Normal uses the pattern rules and structural detection and is the recommended setting. Maximum adds a local name recognition model of about 130 MB that catches names and companies written without any context; it is slower and not offered on phones.

Does it work on scanned documents?

Yes. Scanned pages are read with local text recognition in Spanish and English before detection, and the redaction is applied to the image of the page.

Can I redact something the tool did not detect?

Yes. Search the document for a word to add it as a detection, or use the drawing tools to mark any area by hand with the style you prefer.

What happens to the text on pages that were not redacted?

Nothing. Those pages are copied as they are, so their text stays selectable and searchable.

Protect and Review

All tools

Photos rather than PDFs? Try OléPIC, our image tools.