How to permanently remove text from a PDF (redact, not just hide it)

6 min readpdfmaro

There is a quiet trap in how most people “redact” a PDF. You open the file, grab the highlighter or a black rectangle, draw it over a name or an account number, save, and send it off believing the information is gone. It usually isn't. The text is still sitting in the file, right underneath the black box — and anyone can get it back in seconds.

Why a black box doesn't actually hide anything

A PDF is built in layers. The real text lives in one layer; a shape or highlight you draw sits in a separate layer on top of it. Visually the text disappears, but the characters themselves are untouched. That means someone can:

  • Select and copy underneath the box. Drag across the blacked-out area, copy, paste into a text editor — and the hidden words appear.
  • Delete the shape. Open the PDF in an editor, remove the rectangle annotation, and the text is revealed as if nothing happened.
  • Search for it. If they already suspect what's under there, a find command will still match the hidden text.

This isn't a rare edge case. Governments, courts and large companies have all leaked sensitive information this exact way — publishing documents with black bars that could be copied straight out. The lesson is simple: covering text is not the same as removing it.

What real redaction has to do

To truly redact, you have to destroy the underlying content, not paint over it. The most reliable way to guarantee that in a browser is to flatten the page: render it to a flat image, paint solid black over the areas you're hiding, and rebuild the PDF from that image. Once the page is a single flattened picture, there is no separate text layer left to copy, delete or search. The blacked-out pixels are the only thing there.

That's exactly how our Redact PDF tool works. You open the PDF, drag a box over anything sensitive on any page, and when you download it, each page is saved as a flattened image with your boxes burned in. The text that was under the box is genuinely gone from the file — not hidden, removed.

The honest trade-off

Flattening is what makes the redaction safe, but it changes the file, and you should know how before you rely on it:

  • The whole page becomes an image. After redaction the remaining text is no longer selectable or searchable, because the page is now a picture rather than live text. That's the price of certainty — and for a document you're sending out to hide something, it's usually the right price.
  • File size can grow. Image-based pages are often larger than text-based ones, especially for long documents.
  • It can't be undone. Keep your original. The redacted copy is deliberately one-way.

Don't forget the data you can't see

Redacting the visible text is the main event, but a PDF can also carry information you aren't looking at: document properties like author and title, and sometimes earlier revisions. If a document is going somewhere truly sensitive, it's worth also checking and clearing its metadata separately rather than assuming the black boxes covered everything.

A note on privacy

Redaction is something you usually do because a document is sensitive, so uploading it to a stranger's server to hide information is a bit of a contradiction. Redact PDF runs entirely in your browser — the file is never uploaded, and the flattening happens on your own device. For contracts, medical records and financial statements, that matters.

The short version

  • A black box or highlight only covers text — the real words stay in the file.
  • Anyone can copy, delete the box, or search to recover “hidden” text.
  • True redaction flattens the page so there's no text layer left to recover.
  • The trade-off: redacted pages become images and are no longer selectable.
  • Also clear document metadata for anything highly sensitive.

Need to black something out properly? Open Redact PDF — it flattens the page so the text is truly removed, it's free, and your file never leaves your device.

The tool from this guide

Redact PDFBlack out sensitive text