How to Redact a PDF So It Can't Be Undone (Permanently)

Here is the mistake that gets people in trouble: they draw a black box over a Social Security number, export, and send the file. The document looks redacted. The number is still in there.

This guide explains the difference between covering text and redacting it, how to check whether a PDF is genuinely safe to share, and how to do it correctly — permanently.

Covering is not redacting

A PDF can contain several kinds of “black box”:

The first three are visual edits. The fourth is a content change. Only the fourth survives someone opening the file in an editor.

If you can select, copy, or find the text inside the blacked-out region, you have not redacted anything — you have decorated it.

How to check a redacted PDF in 30 seconds

Before you share any redacted file, run these checks:

1. Try to select inside the black box. Drag your cursor across it. If text highlights or copies, it was only covered.
2. Search for the sensitive string. Use your reader’s Find (Ctrl/Cmd+F) and type part of the SSN, account number, or name. A match inside the box means the text survives.
3. Check for removable objects. Open the file in an editor and inspect the black region. A selectable rectangle that can be deleted is an overlay, not a redaction.
4. Inspect the metadata too. Author names, titles, and tracked changes live outside the visible page — redaction of visible content does not touch them automatically.

Any failure in steps 1–3 means the file is not safe to send.

How to redact permanently

The reliable method burns the redaction into the page pixels rather than adding an object:

1. Load the PDF in a redaction tool that renders pages to Canvas.
2. Mark the sensitive regions — drag boxes over the text you need removed.
3. Use find-and-redact for repeated values. Type the SSN or account number once and redact every occurrence across all pages. Manually missing one occurrence in a 40-page document is the failure that actually happens.
4. Export. The tool rebuilds the PDF from the rendered, redacted pixels — the original characters are not carried into the output.

Because the export is generated from burned-in pixels, there is no underlying text layer left to recover. The black region is the content now.

The two properties that matter in a tool

• Permanence: does it rebuild from pixels, or draw an overlay? (Above: only pixel-rebuild is safe.)
• Where the file goes: a permanent redaction on a server that retains your document is a different risk profile than one done locally.

That second point is why BlurPDF Online runs entirely in your browser: the file is rendered, redacted, and rebuilt on your device, and you can confirm it by watching DevTools → Network show zero file requests. We covered that trade-off in is it safe to upload a PDF to an online tool?.

Common redaction mistakes

• Rasterizing “just in case” without checking the output. Some tools downsample but keep an extractable text layer. Verify with Find.
• Redacting the visible text but not the annotations. Comments and highlights hold their own text.
• Forgetting document properties. Author, company, and title fields leak just as readily as page content.
• Trusting a screenshot of a redacted region. A screenshot pasted over text still leaves the original beneath it.
• Not checking page-by-page on long files. Find-and-redact exists precisely because manual review misses occurrences.

Do it now

Take any PDF with a number you would not want shared, redact it in the browser, then run the three checks above — select, find, inspect. If all three come back clean, you have actually redacted the document.