Why black boxes fail — and what real PDF redaction actually does

Drawing a black rectangle over text in a PDF usually does not delete it. Here is why drawn boxes are recoverable, why that matters, and what true redaction does instead.

It is one of the most common privacy mistakes people make with documents. You have a PDF with some sensitive text — a transaction, a name, a number — so you open it in a viewer, draw a solid black rectangle over the words, and save. It looks completely hidden on screen. So it must be hidden, right?

Often, it is not. And the gap between "looks hidden" and "is hidden" is exactly where redaction fails.

PDFs are not pictures

The key thing to understand is that a PDF is not a flat image. It is a structured document that contains the actual text as selectable, searchable characters, plus a separate set of instructions for how to draw everything on the page. When you draw a black box over some words, you are adding a new drawing instruction on top — a filled rectangle — that visually covers the text. But the original text instruction is still in the file, sitting underneath.

That means the text is still there, fully intact, and trivially readable. Select the area with your cursor and paste into a text editor. Use a copy-all-text option. Run a basic text-extraction tool. Any of these will hand back the words you thought you had hidden, because the box never touched them.

Worse: boxes can be removed

Because the black rectangle is just another object layered on top, it can be peeled off. Move or delete that rectangle — in many editors, a couple of clicks — and the text underneath reappears, perfectly legible. A redaction that can be undone by the recipient is not a redaction at all. It is decoration.

This is not theoretical. There is a long history of organisations — governments, law firms, companies — releasing redacted PDFs where the sensitive text was fully recoverable because someone had drawn boxes instead of removing the content. It is a well-known, well-documented failure mode.

What true redaction does

Proper redaction does not cover the text; it removes it. Using PDF redaction annotations, the tool marks the regions to be redacted and then, crucially, applies them — which deletes the underlying text, vector graphics, and images in those regions from the content stream entirely. After that, the words are gone. Copy and paste returns nothing. Text extraction returns nothing. There is nothing underneath to recover, because it was destroyed.

That is the difference between a box and a redaction. A box says do not look here. A redaction says there is nothing here. Only the second one is safe to send.

Why we do it the hard way

Destroying text takes a fraction of a second longer than drawing a box, and it is what makes the result trustworthy. Every transaction you do not whitelist is removed with proper redaction annotations and then applied, so the underlying text is genuinely gone before the file ever leaves. The output is a PDF you can share knowing the redacted content cannot be recovered by anyone, with any tool.

A note on when to redact at all

Redaction is the right tool when the recipient only needs part of a document — a rental application, an expense claim, an affordability check. It is not right everywhere. Mortgage underwriters and UK visa applications generally require unredacted statements in full, and a redacted file will be rejected outright. Match the tool to the situation.