How to redact a PDF so the text cannot be recovered

Redaction is the operation people most often get wrong, and the mistake is invisible until it matters. Somebody opens a PDF, draws a filled black rectangle over a name or an account number, saves it, and sends it out believing the information is gone. It is not gone. It is underneath.

A PDF is not a picture of a page. It is a list of drawing instructions, and text is stored as instructions that say "place these characters, in this font, at these coordinates". Drawing a rectangle adds one more instruction to that list. It does not delete the ones already there. The text is still in the file, still selectable, still searchable, and still extractable by anyone who copies it out or opens the file with any PDF library.

Why this keeps happening to organisations that should know better

Court filings, government disclosures and corporate reports have all leaked this way, repeatedly, for years. The reason it persists is that the failure is silent. The file looks correct on screen, it prints correctly, and it passes a visual review. Nothing in the workflow tells you the information is still there, so the mistake is only discovered when somebody else finds it.

A second version of the same mistake is changing the text colour to white. The characters are still in the content stream; they are simply drawn in a colour that matches the page. Selecting the area reveals them immediately.

What proper redaction does instead

Real redaction removes the drawing instructions themselves. The characters are taken out of the page’s content stream, so there is nothing left to select, copy or search, and no amount of inspecting the file will bring them back.

This is what PDF Clay does when you delete text. It marks the region, removes the glyphs from the content stream, and only then draws anything new. The original characters are not covered. They stop existing in the file.

There is a detail worth understanding, because it changes what you should expect. When the text sits on an ordinary page, the area underneath is left completely alone — backgrounds, rules and images all survive, because only the characters were removed. When the text sits on top of a scanned image, the words are part of the picture rather than separate instructions, so removing the text layer alone would leave them visible. In that case the area is painted with the colour sampled from the surrounding paper, so the result still matches the page.

How to redact a PDF here

1. 1Upload the PDF and wait for the editor to load the page.
2. 2Click directly on the text you want removed. It becomes editable in place.
3. 3Select it and delete it. The characters are removed from the file, not covered over.
4. 4If you want a visible black bar as well, use Add shape, draw a rectangle over the area and set its fill to black. This is cosmetic; the removal already happened.
5. 5Download the PDF and verify: try to select the redacted area and paste it somewhere. Nothing should come out.

Things that are still worth checking

Removing visible text does not touch everything a PDF can carry. Before sending a sensitive document, consider:

• Document metadata, which can hold the author’s name, the software used and the original filename.
• Earlier revisions. Some editors append changes rather than rewriting the file, leaving previous versions inside it. Exporting a fresh copy avoids this.
• Attachments and embedded files, which are stored separately from the page content.
• Anything in the file name itself, which is often the most overlooked leak of all.

For genuinely sensitive material — legal disclosure, medical records, anything with a regulatory obligation attached — verify the result rather than assuming it. Open the finished file, select everything, and read what you actually get.