How to Permanently Redact a PDF

Draw a black rectangle over text in a normal PDF editor or annotation tool, and the text underneath is untouched — it's still sitting in the PDF's content stream, layered under the box you drew. Anyone can recover it by selecting and copy-pasting the text, or running the file through any text-extraction tool. The box is purely visual, added as a separate object on top of the page; it doesn't touch the text layer at all. This is one of the most common mistakes people make when they think they've redacted a document: viewed in a PDF reader, the black box looks final, but the underlying text is exactly as extractable as it was before. If the file ever gets forwarded, uploaded, or opened in a tool that ignores the annotation layer, the "redacted" information comes right back out. That's the risk this article, and the redaction workflow below, both exist to close.

Real redaction has to remove the text, not cover it. pixeltools' redact-pdf tool at /redact-pdf does this by rasterizing — re-rendering as a flat image — any page that has at least one redaction region drawn on it. Once a page is rasterized, there's no text layer left to extract; what's under the black box is gone, not hidden. To use it: open a PDF, draw a box over whatever needs to disappear (names, account numbers, medical record numbers, signatures), and export. Every page you drew a redaction on gets flattened to an image on export; pages you didn't touch pass through untouched, keeping their original selectable, searchable text. The whole process runs in your browser — nothing is uploaded to a server. You can verify it yourself afterward by trying to select or copy text under the black box in the exported file: there won't be any. Redacted pages are rasterized at 2x the PDF's native resolution and re-embedded as JPEG at 85% quality, the same settings the tool's editor preview uses, so the flattened page stays sharp for reading or printing even though it's now a flat image instead of text.

Redacting a page removes all of its text-selectability, not just the part under the box. Once a page is rasterized, that entire page becomes an image — so any text on that page that wasn't touched by a redaction also stops being selectable or searchable. That's expected behavior, not a bug: rasterizing is what makes the redaction permanent, and it can't be done selectively within a page. It only affects pages you actually redact — a 20-page contract where you redact one signature line only flattens that one page; the other 19 keep their original text layer, fully searchable and copyable. If keeping every page's text searchable matters more than redacting a specific one, that's a genuine tradeoff to plan around: know before you export which pages will lose selectable text.

Redaction matters most anywhere a document has to leave your hands with someone else's private information stripped out. Common cases: legal and compliance documents where names or case details can't go to an outside party; medical or personal records, blacking out patient identifiers before sharing files for research, insurance claims, or education; contracts and financial statements where account numbers or figures need to stay hidden; and FOIA or public-records requests, where the law requires removing exempt information before release. In every one of these, a visual-only black box is a liability, not a safeguard — if the underlying text is still extractable, the document hasn't actually been redacted, it just looks redacted. The stakes are exactly why "it stays under the box" and "it's actually gone" need to be two different, verifiable things, not a detail nobody checks until it's already been sent.

pixeltools' redact-pdf tool is free with unlimited redactions — no page or file caps, no daily quota. That's a real point of difference: Xodo, one of the more common PDF tools, limits its free tier to one redacted PDF per day, which doesn't work if you're clearing a batch of records at once. Because the tool runs entirely client-side, there's also no upload step to worry about — the file never leaves your browser, which matters for anything sensitive enough to need redacting in the first place. There's no account, no watermark, and no limit on how many pages or files you redact.

Try to select the text under the black box, or copy the page's text into another document. If a page went through the redact-pdf tool with a redaction drawn on it, there will be nothing to select — the page is an image now, not text. If you can still highlight or copy anything under or near the box, the file wasn't really redacted, only annotated, and the underlying text is still there for anyone who opens it. This is a quick, low-effort check worth doing before sending any redacted document out: open the exported file, click into the redacted area, and try Ctrl+A / Cmd+A or copy-paste. A genuinely redacted page returns nothing. It's the same check worth running on a document someone else sent you, too, before assuming their redaction actually held.