How to redact a PDF properly, and why a black box is not enough
Every few months a court filing or government report makes headlines because the blacked-out passages could simply be copied out. The black boxes were real, but the text underneath was still there.
Updated:
Why a black rectangle does not hide anything
A PDF is built in layers of objects. When you draw a black shape over a line of text, you add a new object on top. The text object below remains in the file, complete with its characters.
Anyone can select the area and copy it, search the document for a name, or open the file in an editor and move the box aside. Highlighting in black, setting the text colour to black on a black background, or using a comment annotation has the same weakness.
What real redaction has to do
Proper redaction removes the information instead of covering it. The text, the image area and anything else below the marked region must no longer exist in the saved file.
Hidden places matter as well. Document properties can contain an author name, earlier versions can survive incremental saves, attachments and form fields carry their own data, and scanned pages often have an invisible text layer created by text recognition.
How UseTool redacts
The PDF redaction tool first suggests sensitive details such as email addresses or phone numbers, and you decide which areas to black out. Pages with at least one redaction are converted into images, and the black areas are burnt into their pixels. The original text of those pages is not carried into the new file.
Pages without redactions are copied unchanged, so their text stays selectable. The rebuilt document leaves out the original metadata, attachments and active document actions.
- Open the PDF and review the suggested details.
- Add boxes for everything else that must disappear, including signatures and stamps.
- Export the redacted copy. The original stays untouched on your device.
- Test the result before sending it, as described below.
Check the redacted file
Open the new PDF and search for a word you redacted. There should be no match. Try to select text on a redacted page; there should be nothing to select. Copying the whole page into a text editor is a quick extra test.
You can also run the file through PDF to text. If a redacted name appears in the extracted text, something was missed.
The trade-off to accept
Because redacted pages become images, their remaining text can no longer be searched or copied, and the file may grow. That is the price of certainty. For documents shared with courts, authorities or the public, a slightly larger file is far better than a leak.