Black out text in a PDF and have it checked. Drawing a rectangle over words — which is what most redaction tools do — leaves the words sitting underneath, selectable and extractable. This destroys the text layer on the pages you mark, then re-opens the finished document and searches it for everything that was under a box, so the claim is verified rather than asserted.
Step by step
01Add the document
Drop a PDF anywhere on the page. It is read into this tab and never uploaded, and your original file is not modified.
02Drag over what must go
Draw a box across each name, account number or signature. Move between pages with the arrows; hover a box to remove it. Marks are stored relative to the page, so they stay put at any window size.
03Apply and verify
Marked pages are rasterised, the boxes are painted onto the pixels, and the document is rebuilt. Then the finished file is re-opened and its text searched for every run that was under a box.
04Read the verdict, then download
Checks says how many runs were removed and whether any survived. If anything is still readable it says so and names it, rather than reporting success.
Worked example
What actually happens to the file, and what it costs.
Given
The verdict after Apply
Verified removed 7 text runs, none survive
Pages rasterised 2 of 6
4 copied through, text intact
Size 412 KB (was 168 KB)The two pages you marked became images and grew the file; the four you did not are untouched and still searchable. That is the trade, and it lands only where you asked for it.
Why this one
Drawing a black box is not redaction
A rectangle in a PDF is a drawing instruction, and drawing over text does not delete text. Select the area and it copies. Run a text extractor and it prints. Point a screen reader at it and it reads out. Governments, courts and law firms have all published documents redacted this way, and the names came straight back out.
The only certain fix is to destroy the layer
Deleting individual text objects sounds tidier, but a PDF's content stream can carry the same characters in overlapping objects, clipped regions and embedded fonts, so a surgical delete is easy to get subtly wrong. Rasterising the page removes every text object at once — there is nothing left to leak because there is no text layer at all.
Verification, not assurance
After the file is written it is opened again and its text layer is read back, then searched for every run that sat under a box. A tool that merely promises removal cannot tell you it worked; this one looks. If something survived, it says so and names it.
Only marked pages pay the cost
A rasterised page stops being selectable or searchable and usually gets bigger. So pages with no marks are copied through whole, keeping their text, links and bookmarks. Redacting one page of a sixty-page report costs you one page's searchability, not sixty.
Metadata is a separate problem
This removes what is printed on the page. A PDF also carries author and producer strings, revision history, embedded attachments and sometimes earlier versions of the content — none of which a visual redaction touches. Treat them as a separate step before the document leaves your hands.
The judgement call
Rasterising is a blunt instrument. It is the right one in some cases and the wrong one in others.
A document you are about to send or publish
Yes
This is exactly the case redaction failures happen in. The words are destroyed and the result is verified, which no drawing-based tool can claim.
Hiding a signature or a photo
Yes
Images are part of the page raster too, so a box over a signature removes the pixels rather than layering something on top of them.
A long document that must stay searchable
Mark sparingly
Every marked page loses its whole text layer, not just the redacted part. Mark only the pages that need it — the rest keep their text, links and bookmarks.
Removing metadata or hidden layers
Not this tool
Redaction here covers the visible page. Author names, revision history and attachments live elsewhere in the file; strip those separately before sending.
Court filings under a strict protocol
Check the rules
Some jurisdictions specify a particular process or software for redaction, and a few require the redacted text to be replaced by a labelled box. Meet the rule you are given.
Reference
FAQ
Because the box is a drawing and the text is still there. A PDF stores text as objects with positions; painting a filled rectangle at the same coordinates hides it visually and changes nothing underneath. Selecting the area, copying it, or running any text extractor brings the words straight back.
Pages you mark are rendered to an image, the black boxes are painted onto those pixels, and the page is rebuilt from the image. Because the page is now a picture, every text object on it is gone — including the ones under your boxes. Then the finished file is re-read and searched to confirm it.
The pages you marked will not be — they are images. Every page you did not mark keeps its text layer, its links and its bookmarks, because those pages are copied through untouched. If searchability across the whole document matters more than this, mark as few pages as you can.
No. It is read into this page, rasterised, painted and rebuilt on your own machine, and the verification pass runs here too. Nothing is sent anywhere, there is no account, and it keeps working with the network disconnected — which for a document you are redacting is the entire point.
No, and it is important not to assume otherwise. Redaction here covers what is printed on the page. Document metadata, embedded attachments and revision history sit elsewhere in the file and need handling separately before you share it.
150dpi is the default: it reads cleanly on screen and prints acceptably. Choose 300 if the document will be printed and inspected closely, at the cost of a considerably larger file. Below about 120 the type starts to look soft.
Yes. The whole page becomes a raster, so a box over a signature, a photograph or a stamp removes those pixels exactly as it removes text. The verification pass only reports text, because text is the part that can leak invisibly.
Keep going
EXIF viewer & remover
Strip the metadata a visual redaction cannot reach.
Delete PDF pages
Sometimes the right redaction is removing the page entirely.
Extract PDF pages
Send only the pages that should leave your hands.
Compress PDF
Shrink a redacted document whose pages are now images.