How to redact a PDF so the words are really gone
Why a black box is not enough
A PDF page is a list of drawing instructions: put these letters here in this font, fill this rectangle there, draw this picture at this size. Drawing a black rectangle over a name adds one more instruction to the end of the list. The instructions that draw the name are still there underneath it, so the name is still in the file.
Anyone who receives that file can select the hidden words and copy them, search for them, or open it in an editor and move the box out of the way. Court filings, government reports and company documents have been published like this, their blacked-out passages readable by anyone who tried.
Real redaction removes what is under the box, and only then draws the box.
Where else removed words can hide
Covering the text is the best-known way a redaction fails, but not the only one:
- Earlier revisions. Many editors save a change by adding it to the end of the file and leaving the original version intact before it. A redaction saved that way can be undone by reading the earlier revision.
- Pictures. A scanned page is one image. A box drawn over part of it leaves the whole image underneath.
- Recognised text. After OCR, a scanned page carries a layer of invisible text over the image. Covering the image does not remove the layer.
- Comments, bookmarks, form fields and the document's properties. They are stored apart from the page, and can repeat a name or a number you removed from it.
How to redact a PDF with OnePDF
- Open Redact PDF and choose your file. It opens in your browser, on your own device.
- Drag a box over each name, number, signature or picture you want gone. Every box is listed in the side panel.
- Go through the list and take off any box you drew by mistake.
- Apply. The text, pictures and drawings under each box are deleted from the page, cut exactly at the box's edge when the box covers only part of them, along with any comment inside it. A black box is drawn where they were.
- Save. The file is rebuilt rather than added to, so no earlier revision still holds what you removed.
A digitally signed PDF changes when it is redacted, so its signature will no longer match. OnePDF asks before redacting a signed file.
How to check a redacted PDF
Do not trust how it looks. Before you send it, open the saved copy and:
- Select the area around each redaction, copy it and paste it into a text editor. Nothing you removed should paste.
- Search the file for a word you removed, such as a name or an account number. It should not be found.
- Convert it with PDF to Text. The text file holds every word the PDF still has as text, in reading order, with no boxes in the way.
- Check the comments, bookmarks and the document's properties in your PDF reader for anything you removed from the page.
Redact private documents privately
The documents that need redacting are the ones that should not be uploaded anywhere: medical records, legal papers, statements with account numbers. OnePDF redacts in your browser, so signed out the file never leaves your device.
Signed in with a TensorActivity account, a file you open is saved to your account as it is, unredacted, until you save the redacted copy over it. To keep an original out of your account, redact signed out, or choose Keep on this device only for that file.
If the redacted copy should be locked as well, add a password after redacting.