SantaPDFTools

5 min read · SantaPDFTools Team

How to Actually Redact a PDF (Not Just Draw a Black Box Over It)

A surprising number of real-world data leaks happened because someone "redacted" a document by drawing a black box over text in an image editor or a PDF annotation layer — the box sits on top of the text, but the text itself is still there underneath, selectable, copyable, and recoverable by anyone who knows to try. Proper redaction has to remove the underlying data, not just hide it visually.

Why a black box isn't redaction

PDFs store text as actual text data, separate from how it's visually rendered. A rectangle drawn on top, even one that looks solid black, is a separate layer sitting above the text — the text layer underneath is untouched. Anyone can select the "hidden" text and copy it out, or in some cases even see it directly by opening the file in a different viewer that doesn't render the annotation the same way. This exact mistake has exposed social security numbers, financial details, and confidential case information in real published documents.

What real redaction does instead

Redact PDF permanently removes the selected text and any images in the marked area from the document's underlying data, then replaces it with a solid black region — so there's nothing left to select, copy, or extract, because the information is actually gone, not just covered.

How to use it correctly

Double-checking your work

Before sending a redacted document out, it's worth confirming the sensitive text is genuinely gone rather than trusting the process blindly. Compare PDF can highlight every difference between the original and the redacted version, which is a fast way to visually confirm exactly what changed and that nothing sensitive was missed.

Also worth checking: document metadata (author name, comments, revision history) can sometimes contain sensitive information too, separate from the visible page content. If that matters for your document, check the file's properties before sharing, not just the visible text.

← Back to the blog