Open your PDF
Drop the file into the editor.
Covering something with a black box does not delete it. The text stays in the file, one copy-paste away. This removes it.
Covering something with a black box does not delete it. The text stays in the file, one copy-paste away. This removes it.
Drop the file into the editor.
Pick the redact tool and drag across the text. Mark as many regions as you need, across any pages.
Applying rewrites the file. Copy the area afterwards and paste it somewhere: nothing should appear.
Everything you need is ready in the browser.
Drop the file into the editor.
Pick the redact tool and drag across the text. Mark as many regions as you need, across any pages.
Applying rewrites the file. Copy the area afterwards and paste it somewhere: nothing should appear.
Because most tools do not redact. They draw.
A PDF page is not a picture. It is a list of instructions: put this word at this position, in this font, at this size. When a tool "redacts" by drawing a filled black rectangle, it adds one more instruction to that list. The instruction that draws the text underneath is still there, unchanged, sitting beneath the rectangle.
The result looks perfect. The page shows a black bar. But select the area, copy, and paste into a text editor, and the text comes straight back out. A command-line tool like pdftotext will read it without you even having to click. So will anything that indexes documents.
The fix is not a darker box or a flattened copy. It is removing the instruction that draws the text.
Whiteout covers. Redaction removes. They look almost identical on screen and behave completely differently.
| Whiteout | True redaction | |
|---|---|---|
| What it does | Draws over the text | Deletes the text |
| Text still in the file | Yes | No |
| Recoverable by copy-paste | Yes | No |
| Found by a text extractor | Yes | No |
| Page stays searchable | Yes | No, the page becomes an image |
| Safe for confidential data | No | Yes |
Whiteout has a legitimate use: tidying a layout, hiding a stray mark, covering something that was never sensitive. It is the wrong tool for a salary, an address, a medical detail, or a name.
The awkward part is that whiteout is often easier to reach and looks like it worked. If you are unsure which one you used, check with the copy-paste test below. It takes ten seconds and it is definitive.
Open the finished file, select the area you redacted, copy it, and paste into any text editor. If nothing appears, the text is gone.
That test is worth doing every time, on the file you are about to send rather than the one on screen in the editor. A tool showing you a black bar is showing you its own preview. The file is what gets emailed.
If any of those bring the text back, what you did was whiteout.
The pages you redacted become images. That is the trade, and it is worth understanding before you start.
To guarantee text is gone rather than hidden, the affected page is rendered and stored as a picture of itself. Nothing can be recovered from it because there is nothing left to recover, but text on that page can no longer be searched, selected, or edited either.
**Only the pages you touched.** A ninety-page contract with one redaction keeps eighty-nine pages fully searchable. On a one-page document, it is the whole thing.
That trade is what makes the removal real. Any approach that leaves the page fully editable has, by definition, left something behind.
Redact when the page has to stay and part of it has to go: a contract where the fee is confidential, a report where one name is protected, a form where an account number appears in the middle of text that matters.
Delete the page instead when the whole page is the problem. It is simpler, it keeps the rest of the document searchable, and there is nothing to verify afterwards. Removing a page and renumbering takes seconds.
And if the document is going somewhere genuinely adversarial, a regulator, a court, an opposing party, consider producing a fresh document that never contained the information, rather than one it was removed from. Redaction is reliable, but a document that never held the data has no failure mode at all.
It does not redact images. If a name appears inside a scanned page or a photograph, drawing over it will not remove it, because there is no text to remove. It is pixels, and the pixels underneath survive. For that you need to edit the image itself before it goes into the PDF.
It does not find sensitive data for you. There is no pattern search that highlights every account number or email address on your behalf. You mark what should go. On a long document, that means reading it.
It does not remove metadata. A PDF can carry an author name, a file path, and revision history in fields you never see on the page. Redacting the visible text leaves all of that untouched. If that matters, check the document properties separately.
Open the editor to work on your file in the browser, with no account or watermark.
No. Once applied and saved, the text is gone from that file. Keep the original if you might need it.
You can cover an area, but there is no text underneath to remove, a scan is an image. Covering it in a scan does hide it permanently once the page is flattened, which is the one case where drawing over something is genuinely enough.
No. The editor reads the file in your browser and the redaction happens there. It is only stored if you deliberately save it to a project.
Yes. The affected pages are images at print resolution, so they look the same on paper.
Yes. Mark regions across as many pages as you need and apply them together. Pages you did not mark are untouched.
Everything you need to take a document from raw file to reviewed, signed and filed.