How to redact a PDF — and the visual vs. cryptographic redaction trap

Redaction sounds simple — draw black rectangles over the parts you want to hide, save the file, share it. Most "redaction" tools work exactly this way. The problem is that most of those tools don't actually remove the underlying text. The black rectangles cover the text visually, but the original characters are still in the PDF file underneath. Anyone with a PDF editor — Adobe Acrobat Pro, the free LibreOffice Draw, even copy-paste in some viewers — can recover what was supposedly hidden.

The two kinds of redaction

Visual redaction draws opaque shapes (usually black rectangles) on top of the page content. The text underneath is still in the PDF's content stream; the rectangles just cover it visually. Every PDF viewer renders the same way — you see the black bars. But the text-layer is intact: copy-paste from the rendered page might still get you the original text (depends on the viewer), and any PDF editor with content-stream access can remove the rectangles and recover everything.

Cryptographic / content-stream redaction actually removes the text from the PDF. It walks the content-rendering operations on the page, identifies which characters fall inside the redacted regions, removes them from the rendering instructions, removes them from the text-extraction layer, and removes them from the document's search index. After this kind of redaction, the original text is genuinely gone — recovering it from the file is not possible.

The two operations produce visually identical results. A recipient looking at the page sees black bars either way. The difference is in what's underneath. For handling decisions on sensitive documents, that difference is enormous.

When visual redaction is fine

Screenshots and printed copies. Once rendered to pixels, the original text is genuinely gone (the pixel data is just black). Anything that goes to print or to a screenshot is permanently redacted.
Sharing with non-technical recipients. A recipient who opens the PDF in Acrobat Reader and reads it sees only the black bars. They'd need to know about PDF editors and want to do extraction work to recover the text.
Quick preview-quality redactions. Sending a sanitized preview while preparing a properly-redacted final version. Use visual for the preview, cryptographic for the final.

When visual redaction is dangerous

Public web posting. A determined reader can run the PDF through forensic tools and recover the text. There have been multiple high-profile incidents where redacted PDFs posted publicly had their redactions defeated within hours.
Legal disclosure (FOIA responses, court filings, discovery production). Visual redaction is not legally sufficient for these contexts. Several agencies have faced lawsuits over recoverable redactions.
National security / classified material. Visual redaction is explicitly disallowed by most classification handling rules.
Anything sent to a journalist, researcher, or technically-capable adversary. Assume they know about and will check for recoverable redaction.

What our tool does — honestly

Our PDF Redaction tool does visual redaction only. We're explicit about this on the tool page and again here. The tool draws opaque black rectangles over the regions you mark using pdf-lib's drawRectangle primitive. The underlying text and image data are still in the saved PDF — covered, but recoverable by someone with the right tool.

We chose this tradeoff explicitly. Doing real content-stream redaction in the browser would require running text-layout-aware logic on every redacted region, identifying which characters fall inside the rectangle, removing them from the page's text-rendering operations, removing them from the page's text-extraction stream, and rewriting the document's search index. pdf-lib doesn't expose those primitives. Doing it correctly requires either a commercial library or Adobe Acrobat Pro. We'd rather ship an honestly-labeled visual redaction tool than a "redaction" tool that gives users a false sense of security.

Step-by-step

Open justfiletools.com/tools/pdf-redaction.
Drop a PDF. The tool renders each page as a thumbnail you can draw on.
Navigate page by page with the arrow buttons. On each page, click and drag to draw black rectangles over the content you want to hide.
Hover over a drawn rectangle to remove it (trash icon). Use "Clear page" to wipe all rectangles on the current page.
Once all redactions are placed, click Apply redactions and save. The output PDF has opaque black bars baked into the page content.
Open the output and verify the black bars cover what you wanted before sharing.

If you need real content-stream redaction

Use Adobe Acrobat Pro's Redact tool. It actually walks the content stream and removes the underlying text:

Open the PDF in Acrobat Pro.
Tools → Redact.
Select text to redact. Acrobat highlights the regions.
Click "Apply" — Acrobat warns that redactions are permanent.
Save the redacted PDF.

Acrobat's redaction is the gold standard. It also has a "Sanitize Document" feature that strips other forensic remnants (form data, hidden layers, metadata, scripts). For high-stakes disclosure, run Sanitize after applying redactions.

Alternative open-source approaches

qpdf with content stream rewriting. qpdf is a free command-line PDF toolkit. You can rewrite the content stream of pages to remove text, but doing this correctly requires understanding the PDF spec.
"Convert to image PDF" + visual redaction. Run the PDF through PDF to Image (rasterizes every page), then reassemble via Image to PDF. The output has no text layer at all — every page is just pixels. Then add visual redactions. This genuinely removes the text (because there's no text left) but the file size is dramatically larger and search/copy stops working entirely.
Print to PDF after visual redaction. Print the visually-redacted PDF to a new PDF (most OSes have a "Print to PDF" option). The print process rasterizes content, removing the text layer. Slower than the image-conversion approach but works on every OS without extra tools.

Common pitfalls

Trusting visual redaction in legal contexts. Multiple agencies have been embarrassed by recoverable redactions. The Manafort court filing in 2019 redacted text with black rectangles whose underlying text was readable through copy-paste. Don't be that example.

Forgetting metadata. The PDF's author, creator, and modification-date metadata can also leak sensitive information. Run the output through PDF Metadata Viewer with the Strip option after redacting.

Forgetting embedded thumbnails. Some PDFs include a thumbnail image of each page in the metadata. Visual redaction on the page doesn't remove these thumbnails — the thumbnail might show the un-redacted version. Acrobat's Sanitize removes thumbnails; many other tools don't.

Privacy considerations

Redaction is, by definition, applied to sensitive documents. Using an upload-to-someone's-server redaction service is contradictory — the sensitive document is at the service's hard drive before the redaction is applied. Our in-browser tool resolves this: pdf-lib applies redactions in your tab. Network panel stays empty during use.

The privacy guarantee here is about upload, not recoverability. The redaction itself is visual (recoverable from the file). For non-upload privacy + irrecoverable redaction, use Acrobat Pro locally or rasterize the visually-redacted output.

Related tools and guides

PDF Redaction — the tool this guide covers.
PDF Metadata Viewer — strip identifying metadata after redacting.
PDF to Image — rasterize for irrecoverable redaction.
Image to PDF — reassemble after rasterizing.
PDF Watermark — add CONFIDENTIAL marking on top of redactions.