How to redact a PDF properly
A black rectangle over a name looks like redaction, but in most PDFs the name is still there underneath, ready to be copied out. Here's why it happens and how to remove information for good.
Updated
Why black boxes don't work
A PDF page is built in layers. Text is stored as characters, and anything you draw, whether a rectangle, a highlight or a comment, is added on top. Cover a sentence with a black shape and the characters are still on the page: select across the box, copy and paste, and the text appears. Search finds it, and screen readers read it aloud.
This has caught out law firms and governments. In 2019, a court filing by lawyers for Paul Manafort was published with redactions that anyone could reveal by copying the blacked-out text, and it made headlines.
Methods that leave the text behind
- Black rectangles or shapes drawn in Word, PowerPoint or a PDF editor, then saved as PDF.
- Highlighting text in black.
- Changing the text colour to black on a black background, or white on white.
- Cropping a page: the hidden part of the page is usually still in the file.
- Covering text with an image or a sticky note.
- Saving over the original in an editor that appends changes: the earlier version can remain inside the file.
How to redact properly
Proper redaction removes the content itself, not just its visibility. Dedicated redaction tools either delete the characters and image pixels under each box, or rebuild the page without them.
PDF Workdesk's Redact PDF does the second: each page with a redaction is rebuilt as an image at 200 DPI with the black boxes burned in, so there is no text left underneath to find. You can search for every occurrence of a word or name, or draw boxes over signatures, photos and handwriting. It runs in your browser, so the unredacted document never leaves your device.
Check before you share
- Try to select and copy the redacted area. Nothing should paste.
- Search the document for each redacted word. There should be no results.
- Look at the document properties: the title, author and subject can name people too.
- Check every variant: full names, initials, email addresses, phone numbers with and without spaces, account numbers.
- Scanned pages are images, so search can't find their text. Draw boxes over it by hand.
Tools for this
Frequently asked questions
Can redacted text be recovered?
With proper redaction, no: the content is gone from the file. With a black box drawn on top, often yes, by copying and pasting.
Does printing and scanning a document redact it?
Printing a page with black boxes and scanning it does remove the text, because the scan is only an image. It's slow and loses quality, and you still need to check the metadata.
Is it safe to redact a PDF online?
Most online tools upload the document, which sends the very information you're removing to a third party. PDF Workdesk redacts in your browser, so the file isn't uploaded.