How to redact a PDF so the text is really gone

Last updated: July 7, 2026

There is a mistake that keeps appearing in court filings, government releases and corporate disclosures: someone opens a PDF, draws a neat black rectangle over a name, a salary or an account number, saves the file and sends it out. The document looks redacted. It is not. The text is still there, sitting quietly underneath the rectangle, and anyone who selects that area and presses copy gets the secret pasted straight into their clipboard.

This exact failure has embarrassed law firms, ministries and police departments over and over again. The pattern is always the same: a journalist drags the cursor across the black bars, copies, pastes into a text editor, and the supposedly hidden names appear in full. Nobody hacked anything. The redaction was never real.

Covering is not removing

A PDF page is not a picture. It is a stack of layers: a text layer with the actual character data, image layers, vector shapes, annotations. A black rectangle drawn in a typical PDF editor is just a new shape on top of the stack; the text beneath it is untouched. Every one of these still works on the "hidden" content:

• Select and copy. Drag over the black bar, press copy, and the original text comes out.
• Search. The find function in any PDF reader will happily jump to text under a rectangle.
• Text extraction. Screen readers, indexing tools and one-line scripts pull the full text layer out of the file, rectangles and all.
• Deleting the shape. In an editor, the rectangle is just an object. Select it, delete it, and the page is back to its original state.

The same trap exists in adjacent forms: highlighting text in black, setting text colour to match the background, or placing a white box over a paragraph. All of it survives copy-paste. Real redaction has one test: after saving, the sensitive text must not exist anywhere in the file, in any layer, in any form.

How to redact a PDF in your browser

robinpdf's Redact PDF tool handles this correctly, and it runs entirely on your own device. That matters more here than for any other PDF task: a document that needs redacting contains, by definition, something sensitive. Uploading it to a random server so a stranger's machine can "help" you hide the sensitive part is backwards. Here the file never leaves your browser, so the unredacted original is never in anyone else's hands, not even for a second.

1. Open the Redact PDF page.
2. Click the box and select your PDF, or drag it in.
3. Draw black boxes over every piece of content you want removed: names, figures, signatures, whole paragraphs. You can place as many boxes as you need, across any pages.
4. Check each page once more. The redaction you forgot to draw is the one that ends up in the news.
5. Press Redact PDF and the finished file downloads automatically.

Here is the important part: when you save, the tool does not simply stamp rectangles onto the file. It flattens the pages to images. Each page is re-rendered as a picture of itself with the black boxes burned in, and the original page, including its entire text layer, is discarded. There is no rectangle to delete and no hidden text to copy, because the characters under the boxes no longer exist in the file at all.

Why flattening works, and what it costs

Flattening is a deliberately blunt instrument, and that bluntness is exactly why it is safe. Tools that try to surgically remove individual characters from the PDF's internal structure must get every edge case right: text split across drawing operations, unusual font encodings, content repeated in annotations or form fields. Miss one and the leak survives. Rasterising the page sidesteps all of it: what you see is literally all there is. The trade-offs are real, though:

• The text layer is gone everywhere on the page, not just under the boxes. The redacted file cannot be searched or copied from. For a document being released rather than edited, that is usually acceptable; recipients still read and print it normally.
• The file may grow. Images of pages are often heavier than the compact text instructions they replace. If the result is too large to email, run it through Compress PDF; compressing the redacted version is safe because there is no hidden layer left to preserve.
• Do the redaction last. Flattened pages are awkward to edit afterwards, so merge, reorder or crop first and redact as the final step.

The other leak: metadata

Redacting the page contents is only half the job. Every PDF carries a metadata block most people never look at: author, title, subject, keywords, the creating software, creation and modification dates. This is where redactions quietly fail even when the pages are clean. A supposedly anonymous document whose author field contains a full employee name is not anonymous; a title field reading "Draft settlement, do not circulate" tells its own story. Because robinpdf rebuilds the redacted file from page images, the original's metadata does not ride along, but you should still verify. Open the result in Edit PDF Metadata and clear or rewrite anything that identifies people, machines or internal project names.

And do not forget the filename itself. "smith_termination_v3_FINAL_legal_review.pdf" leaks more than some metadata fields do. Rename the file to something neutral before it goes anywhere.

A pre-sharing checklist

Run through this every time, in order. It takes two minutes and it separates redacted from "redacted":

1. Redact and flatten. Draw boxes over everything sensitive in Redact PDF and save, so pages become images and the hidden text is destroyed.
2. Check the metadata. Open the new file in Edit PDF Metadata and clear or neutralise author, title, subject and keywords.
3. Rename the file. Give it a neutral name that reveals nothing about people, versions or internal context.
4. Verify like an attacker. Open the final PDF, try to select text over every black box, and search for one of the redacted words. Both must come up empty; if anything can be copied out, go back to step one.

Never assume a redaction worked; demonstrate it on the exact file you are about to send. For the wider picture on document tools and privacy, see our guide on whether online PDF tools are safe.

Frequently asked questions

Is drawing a black box in a normal PDF editor ever enough?

No. Unless the tool removes the underlying content and flattens or rewrites the page, the text remains in the file and can be copied out. A box alone is decoration.

How do I check that a redaction actually worked?

Open the final file, drag-select across each black box and try to copy, then use the search function to look for a word you redacted. If nothing can be selected or found, the content is gone.

Does my file get uploaded anywhere?

No. Everything runs in your browser on your own device. The unredacted original never leaves your machine, which is exactly what you want for a document this sensitive.

Will the redacted PDF still be searchable?

No, and that is by design. Flattening removes the entire text layer, which is what guarantees the hidden text cannot be recovered. Keep a searchable original in a safe place and share only the flattened copy.

Why did my PDF get bigger after redacting?

Pages stored as images usually take more space than pages stored as text instructions. If size is a problem, run the redacted file through Compress PDF.

Can a redaction made this way be reversed?

No. The saved file contains only a picture of each page with the boxes burned in: nothing to delete, nothing to read underneath. Keep your original somewhere secure, because the shared copy cannot give the content back to you either.

Redact your PDF now →