Why visual redaction fails
In a PDF, the text layer and the graphics drawn on top are independent. A black rectangle is just another graphic. The words beneath it remain in the content stream, so anyone can select, copy, or extract them — and search engines and scripts can read them instantly.
There is a long history of "redacted" documents leaking exactly this way: the box is there, but copy-pasting reveals the text, or a downstream tool dumps it.
How Scrubbr redacts for real
For each page you mark, Scrubbr renders the page to a high-resolution raster, paints your redaction boxes as solid black onto that image, and rebuilds the page from the flattened image. Because the page is now an image, the original text and vector content under (and around) the boxes no longer exist in the file.
Pages you do not mark are copied through untouched, so their text stays selectable. Marked pages become images — a deliberate trade-off that guarantees removal.
The verification step
After redaction, Scrubbr extracts the text that sat under your boxes in the original, then re-extracts all text from the output and confirms none of it survived. It also confirms each redacted page yields zero extractable text. If anything leaked, it tells you.
How to truly redact a PDF
- 1Open the PDF. Drop your PDF onto Scrubbr; pages render locally for marking.
- 2Draw redaction boxes. Drag rectangles over the text or areas you need removed.
- 3Apply redaction. Scrubbr flattens marked pages so the underlying content is destroyed.
- 4Verify and download. Scrubbr re-extracts text to prove the redacted content is unrecoverable, then you download the result.
Frequently asked questions
Can the redacted text be recovered?
No. Marked pages are rebuilt from a flattened image, so the original text and vector objects are not present in the output. Scrubbr verifies this by re-extracting text.
Will my redacted pages still have selectable text?
Pages you redact become images (not selectable). Pages you leave alone keep their original selectable text.