How to safely redact personal data in a PDF

The biggest mistake: covering instead of removing

The most common method is drawing a black rectangle over the text or painting over it in a graphics program. If the rectangle is just another layer on the page, the text underneath still exists in the file. It is enough to select and copy it, and often it can even be found with the search function of a PDF reader. There are known cases of published documents where someone simply removed the black boxes and read the “redacted” content. Real redaction must permanently remove the data from the file.

How PDFowo does it

The “Redact” tool in the editor marks the area that should disappear. On saving, pages with redactions are turned into images, so the text hidden under the black box no longer exists in the file. The cost of this approach is clear: such a page no longer has a text layer, so text on it can’t be selected or searched. For a document you send to someone to read, that is usually a good compromise. The other pages, without redactions, stay unchanged.

Searching the whole document for data

In a long document it is easy to miss one occurrence of an ID number. In the editor, use search (Ctrl+F) and the “⋯” options panel: there you will find data patterns that the program recognises across the whole file – e-mail addresses, phone numbers, payment card numbers and Polish identifiers (PESEL, NIP, IBAN PL). For numbers that have a checksum, the patterns check it, so random strings of digits are not treated as personal data. You can also redact all the matches found in one click.

A checklist before sending a document

1. Mark what should disappear. Not just numbers: names, addresses, signatures, licence plates, contract numbers.
2. Redact and save the file. Use real redaction, not drawn rectangles.
3. Check the result. Open the saved file, try to select text where the redaction is and use search. Nothing should be found.
4. Take care of the document properties. The author and title in the metadata can give data away. When you save with redactions, PDFowo removes them automatically, together with accessibility tags and form fields from the redacted pages, and replaces the redacted text with █ in bookmarks and notes too.
5. Keep the original separately. Redaction is irreversible, so keep the original in a safe place.

Why it matters under the GDPR

If you share a document containing personal data, you should limit it to what is actually needed. Accidentally revealing hidden data can be a data breach. This is not legal advice, but a practical rule: the less data in the file, the lower the risk. Because PDFowo processes files locally in the browser, the documents themselves are not uploaded to a server during redaction.

Example: a contract passed on to an accountant

You are sending a copy of a contract to your accountant and don’t want to reveal the other party’s account number and ID number. Open the file in the editor, start search (Ctrl+F) and choose the relevant pattern, e.g. bank account numbers. The program highlights every match in the whole document, so you don’t have to go through it page by page. Redact them (manually with the “Redact” tool or all at once in one click), save the file and check the result: try to select text in the redacted spots and search for the number in a PDF reader. If nothing is found, the file is safe to send.

Frequently asked questions

Can a redaction be undone after saving?

No. The data is permanently removed from the saved file, so keep the original separately.

Why can’t I select text on a page with a redaction?

Such a page is saved as an image so that the hidden text really disappears. The other pages, without redactions, keep their text layer.

Will pattern search find all personal data?

No, it is a help, not a guarantee. The patterns recognise typical formats (e-mail, phone, card, Polish PESEL, NIP and account numbers), but data written in an unusual way or in a scan has to be checked yourself. For a scan, use OCR first.

Can I redact part of a scan?

Yes. Redaction works on an area of the page, so it can cover both text and images, e.g. a signature or a photo.