How to Redact a PDF Properly: Black Boxes That Actually Delete

4 min read · Updated 2026-10-27 · PdfPixel Editorial Team

Redaction is the one PDF task where a mistake is not cosmetic. A highlight left in the wrong colour annoys nobody; a redaction that only looks like a redaction hands your confidential text to anyone curious enough to drag a cursor across the black box. This guide is about the difference between covering information and removing it, and the two-minute ritual that proves which one you did.

A black box is a promise, not a proof

A PDF page is a list of drawing instructions: put this text here, this image there, this rectangle on top. When someone "redacts" by drawing an opaque rectangle over a paragraph, every original instruction stays in the file, merely painted over. The text remains selectable, searchable and copyable. Select all, copy, paste into a text editor, and the confidential paragraph appears under the black paint. This is not a hacker trick; it is copy and paste. Real redaction deletes the content instructions inside the marked region and only then paints the opaque fill, so what remains is a hole in the page with a lid on it, not a hidden room.

How PdfPixel's redaction works

In the Redact PDF tool you draw boxes over exactly the regions that must disappear: names, numbers, amounts, addresses, case references. On export the marked regions are sanitized: the underlying text and image content inside each box is discarded and replaced with an opaque fill baked into the page. The exported file therefore contains nothing to select, search or copy inside those regions. Because the whole operation runs in your browser, the unredacted original never leaves your device either, which matters when the document is the reason you are being careful in the first place.

The verify-before-sharing ritual

Never share a redacted file on trust. Run three checks on the exported copy, not the master. First, drag a selection across each redacted region: nothing should highlight. Second, search the document for one exact string you removed, a surname, an account number: zero results. Third, extract the file to plain text and skim it, because extraction ignores visuals entirely and shows only what the file truly carries. Three checks, ninety seconds, and the difference between a redaction and a decoration is settled before the recipient discovers it for you.

What people forget to redact

The paragraph was never the only leak. Metadata carries the author name, the editing tool and sometimes the original filename; strip it with the metadata remover before sharing. Headers and footers repeat case numbers and client codes on every page, far from the paragraph you boxed. Bookmarks and internal links can jump a reader straight to a section you thought hidden. Attachments embedded in the PDF travel with it, unredacted. And if the page is a scan, remember the "text" is pixels: your box must fully cover the printed pixels, edge to edge, because a two-millimetre gap is a readable sliver at zoom.

When redaction is the wrong tool

Sometimes the honest answer is withholding, not editing. If a requester is entitled to the full document, a redacted copy may breach the very rule you are trying to honour. If only one clause matters, sending the single relevant page is cleaner than a Swiss-cheese contract. And where a legal hold or a court order governs the document, touch nothing until you know what the order permits: sanitizing the wrong copy can destroy evidence you were required to preserve. Redaction is a scalpel for sharing; it is not a substitute for deciding what should be shared at all.

Frequently asked questions

Every step above runs inside your browser. The unredacted original and the redacted export both stay on your device.