A Black Box Over Text Is Not Redaction

It looks convincing. You open a PDF, draw a black rectangle over a name, an address, or an account number, save the file, and the sensitive text disappears from view. The page looks clean. The problem is that looking clean is not the same as being clean. In most PDF editors, that black rectangle is simply a shape drawn on top of the text. The text underneath is still there, intact, in the file — and anyone who knows where to look can recover it in seconds.

This is the most important distinction in PDF redaction, and the one most often missed. True redaction removes the content. Covering it merely hides it from sight. Redact PDF performs true redaction: pages with marked regions are rebuilt as images, so the covered text is removed rather than hidden, and the work happens in your browser.

Why covered text survives

A PDF is not a photograph of a page. It is a structured file containing text, fonts, positions, and drawing instructions, layered together. When you draw a black rectangle in an editor, you are usually adding a new drawing instruction on top of the existing text layer. The editor does not delete the text beneath, because you did not ask it to delete anything — you asked it to draw a shape. The text remains in the file's content stream, with its characters, its position, and its font, exactly as before.

That is why the classic test is so damning. Open a PDF that has been "redacted" with a black box, select the black area with your mouse, copy it, and paste it into a text editor. Very often, the hidden text appears in the paste, complete and readable. The same applies to searching: a reader's find function can locate a "redacted" name, because the search looks at the text layer, not at what is drawn over it. Removing the rectangle in an editor, or extracting the text layer with a tool, reveals the same content. The black box was never a removal; it was a curtain, and curtains can be pulled aside.

This is not a rare bug in one editor. It is how layered document formats work, and it has caused real disclosures in legal, government, and corporate documents where a black box was mistaken for a redaction. The lesson is blunt: if you can select, copy, or search the text, it has not been redacted, however black the box looks.

What true redaction does differently

True redaction removes the text from the file, not just from view. In Redact PDF's approach, pages containing marked regions are rebuilt as images. That means the page is rendered, the marked areas are removed from the rendering, and the resulting page contains pixels, not a hidden text layer carrying the sensitive words. There is no underlying text left to copy, search, or extract for those regions, because the text is no longer stored there in text form.

That approach has a consequence worth understanding: a rebuilt page's remaining text also becomes image pixels, so it is no longer selectable or searchable on that page. That is a fair trade for a redacted page, where the priority is that nothing sensitive survives, but it means you should redact a copy of the document, not your only original, and keep the original safely stored if you may need its text later.

True redaction also requires care about what else carries information. A document's metadata — title, author, comments, hidden layers — can contain the very details you removed from the visible page. A thorough redaction workflow checks those places too, and does not assume that cleaning the visible page cleans the whole file.

Worked example: a fictional invoice, covered versus redacted

Imagine a fictional invoice, invented for this example, from a fictional supplier to a fictional customer named Alex, at alex@example.org. The invoice contains an account number that must not be shared when the invoice is forwarded as a payment example. In the first version, someone draws a black rectangle over the account number in an editor and saves the file. Viewed normally, the invoice looks fine: a neat black bar where the number was.

Now test it. Select the black bar, copy, and paste into a plain text document. The account number appears in the paste. Search the PDF for part of the number, and the reader jumps straight to the black bar. The "redaction" has failed every test that matters, while passing the only test that does not: looking right.

In the second version, the same invoice is processed with Redact PDF, marking the account number region for true redaction. The marked page is rebuilt as an image, with the number genuinely gone. Select, copy, and search now find nothing in that region, because there is no text there to find. The page looks similar — a dark area where the number was — but the similarity is only visual. Underneath, the two files are fundamentally different: one hides, the other removes.

If the invoice also needs its pages reordered before sharing, Organize PDF can sort them first, and if it needs combining with a covering letter, Merge PDF joins the files — with redaction done as the final step, after all other edits, so no earlier version leaks through.

Honest limitations

True redaction removes content, and removal is irreversible for that copy. If you redact the wrong region, or redact too broadly, the information is gone from that file, and you must return to your unredacted original. Always work on a copy, keep the original safe, and verify the redacted file before sending it anywhere — the verification steps are the subject of a separate post.

Rebuilding a page as an image also means that page loses selectable text, as noted above, and the file size and appearance can change slightly. For most redacted documents, that is acceptable, but it is a reason to redact only the pages that need it, rather than rebuilding an entire long document unnecessarily.

Finally, redaction is only as good as the marking. If you miss an occurrence of a name — in a header, a footer, a repeated table, or a second mention on another page — that occurrence survives, however thoroughly the marked regions were removed. Search the original for every occurrence before you redact, and search the redacted file afterwards to confirm nothing remains.

For the wider workflow of handling PDFs before sharing, see how to merge and split PDFs.

FAQs

Why can blacked-out text in a PDF still be copied?

Because the black rectangle is usually drawn on top of the text layer, not instead of it. The text remains in the file, so selecting, copying, or searching can still recover it. Covering is hiding, not removing.

What is true PDF redaction?

True redaction removes the marked content from the file, so it cannot be selected, copied, searched, or extracted. In Redact PDF, marked pages are rebuilt as images, so no hidden text layer survives in those regions.

How can I test whether a redaction worked?

Try to select and copy the redacted area, search the document for the removed words, and extract the text. If any of those recover the content, the redaction failed. Test the file you intend to send, not a different version.

Should I redact my only copy of a document?

No. Redact a copy, keep the original stored safely, and verify the redacted copy before sending it. Redaction is irreversible for the file it is applied to.

Related reading: How to verify the redaction worked.

human review pending client designation