Redact Text in a PDF Permanently, Not Just Cover It
Redacting a PDF means removing sensitive content so it cannot be recovered, not merely covering it with a black box. In this tool you drag boxes over the passages to hide, and every affected page is then flattened to an image with the boxes burned in, so the underlying text objects are destroyed and cannot be selected, copied, or extracted. Redacted pages become pictures and lose selectable text. Everything runs locally in your browser.
How it works
The redaction failures that make headlines share one mechanism: someone drew an opaque shape over text, and the text object survived underneath. A PDF page is a stack of drawing commands, and painting a black rectangle simply appends one more command — the glyphs beneath it are still in the content stream, still returned by select-all and copy, still found by search, still extracted by any text tool. Court filings and government releases have leaked names and figures exactly this way. Effective redaction must therefore change what the file contains, not what it shows.
This tool takes the bluntest reliable route. First, each page is rendered on screen with pdf.js so you can see the real layout and drag redaction boxes over anything sensitive — a name, an account number, a whole paragraph, part of an image. Rendering for display and marking regions changes nothing yet.
When you apply the redaction, every page that carries at least one box is rendered to a bitmap at high resolution, the boxes are painted onto that bitmap as solid black rectangles, and the page in the output document is replaced by this image. The original content stream, its text operators, its fonts, and its embedded images for that page are discarded entirely — the page is now a photograph of itself with the black bars burned into the pixels. There is no layer to peel back and no string to extract, because no text objects exist on that page anymore.
The trade-off is honest and unavoidable: flattened pages are no longer selectable or searchable, even in their unredacted parts, and they usually take more bytes than the text-and-vector original. Pages without boxes are copied through untouched, keeping their text layer. Two adjacent risks remain outside redaction's scope: document metadata (author, title, timestamps) survives — clean it with Remove PDF metadata — and content merely hidden by cropping elsewhere in the file is a separate hazard covered on the Crop PDF page.
Worked example: releasing a 9-page incident report
An HR team needed to share a 9-page incident report (1.1 MB) with an external reviewer, minus employee names and salary figures. The sensitive strings appeared on pages 2, 3, and 7 — eleven passages in total. In the tool, they dragged eleven boxes across those three pages; pages 1, 4, 5, 6, 8, and 9 were left unmarked.
On apply, the three marked pages were rasterized with the bars burned in, and the output came to 1.6 MB — the growth entirely from three full-page images replacing compact text streams. Verification took thirty seconds: select-all on page 2 selects nothing, searching for a redacted surname returns no hits anywhere, and running the file through a text extractor yields the six untouched pages' text and nothing from the flattened three. Compare that with their first attempt in a drawing tool, where the "redacted" file happily gave up every name to copy-paste.
Frequently asked questions
Why is drawing a black rectangle over text not real redaction?
A drawn rectangle is just another object painted on top of the page, and the text object underneath remains fully intact in the file. Anyone can select the covered area and copy the text out, or open the file in an editor and delete the rectangle. Real redaction removes the content itself, which is what this tool does by flattening the page.
Can redacted text be recovered from the file this tool produces?
No. Redacted pages are rebuilt as images rendered with the black boxes already applied, so the output file contains no text objects for those pages, only pixels. The pixels behind the boxes are solid black and carry no trace of the original glyphs.
Why can I no longer select text on the redacted pages?
Flattening replaces the whole page with a rendered picture, so even the unredacted parts of that page stop being selectable or searchable. This is the deliberate cost of destroying the hidden content. Pages you did not draw boxes on are left untouched and keep their text layer.
Does redacting also remove the PDF author and title metadata?
No. Redaction destroys content inside the pages you marked, while document metadata such as author, title, and timestamps lives elsewhere in the file. Run the result through the Remove PDF metadata tool for a fully clean handout.
Is it safe to redact confidential documents in an online tool?
It is safe here because the rendering and rewriting happen in JavaScript inside your browser, and the document never leaves your device. With cloud redaction services the unredacted original is transmitted to a server, which is itself a disclosure.
Will the redacted file be larger than the original?
Often yes, because a page stored as compact text and vector commands becomes a full-page image after flattening. The increase depends on how many pages you redacted and their rendering resolution. Untouched pages contribute no growth.