Redacting a PDF: the black box hides nothing

The gesture that does not protect

You receive a document to pass on. It contains a name, an amount, an address that are none of the recipient's business. You open the PDF, pick the rectangle tool, fill it black, and drop it on top. On screen the passage is gone. You send it.

The text is still there. You did not delete it, you drew over it. A PDF is not an image: it is a stack of drawing instructions, and you have just added one at the end. The earlier ones have not moved.

Anyone can get it back. Selecting the passage with the mouse and copying it is usually enough — the text copies from under the rectangle, because it is still in the file. A keyword search finds it. Converting to plain text exposes it entirely. None of this requires any particular skill.

Why it happens so often

Because nothing warns you. The document looks right on screen, looks right in print, and looks right to a recipient who is not looking. The flaw only shows to someone who is looking — which is precisely the person you were protecting yourself against.

And because tools offering "highlight" or "annotate" never say that this is not redaction. They do what you asked: they draw.

What real redaction is

Redacting means removing the content from the file, not covering it. Afterwards the text no longer exists: there is nothing to copy, nothing to search, nothing to extract. The black rectangle that remains is a visual mark saying something was there, not a lid placed over it.

That is what our redaction tool does: the area you draw is removed from the page content. Processing happens in your browser, which means the confidential document you are cleaning is not sent anywhere while you clean it — that would rather defeat the point.

Three things redaction does not solve

Removing visible text is necessary, but a PDF carries more than it displays. Before sending a sensitive document, also look at:

• Metadata. Author, software, internal title, sometimes the full path of the file on the workstation — so the user's name, and sometimes a client's name in a folder name.
• Earlier versions. A PDF can contain its own history: successive saves stack up instead of replacing each other. A version from before the correction may still be readable inside the file.
• Attachments and hidden layers. A PDF can embed other files, and layers merely made invisible remain extractable.

Cropping is not redacting

The confusion is worth naming because it mirrors the previous one. Cropping a PDF moves the visible frame of the page; whatever falls outside stays in the file and comes back by widening the frame.

That is perfectly legitimate for trimming white margins or recentring a scan. It is not a way to make an awkward header disappear.

Checking your work

The check takes ten seconds and there is no reason to skip it on a document that commits you.

Open the redacted file, press Ctrl+A (or Cmd+A) to select everything, copy, and paste into a text editor. What appears is what the recipient will see if they do the same. If the redacted passage is on your clipboard, the redaction failed.

Use search too: look inside the PDF for a word you thought you had removed. If it is found, it is still there.