How to Redact a PDF So the Text Is Really Gone

🔏 PDF & privacy⏱️ 7 min readFree tool included

The most common redaction mistake is also the easiest one to make: drawing a black box over the sensitive part and calling it done. The box is just ink on top of a page — the words underneath are still in the file, one copy-paste away from disclosure. Here's how to tell the difference between covered and destroyed, and how to actually destroy.

Advertisement

Why a black box isn't redaction

A PDF page is a set of drawing instructions: put this glyph here, that rectangle there. When you draw a black box over a Social Security number, you've added one more instruction — after the ones that draw the number. Every serious PDF tool still honors the earlier instructions when asked for text: select the region and paste it into a text editor, or run an extractor like pdftotext, and the "redacted" value comes back out. It's not an obscure attack; real court filings, government reports, and corporate documents have leaked names, account numbers, and addresses exactly this way.

The same logic sinks the other shortcuts. A highlighter is formatting, not removal. Deleting text in a basic editor often leaves the object recoverable in the file's structure. Password-protecting the PDF or setting a "no copy" permissions flag restricts well-behaved viewers while removing nothing — any tool that ignores the flag reads everything. Removal means the bytes are gone.

How do you check whether a PDF is really redacted?

Three tests, thirty seconds:

  1. Select test. Click and drag across the redacted area. If an invisible text cursor grabs something, there's still text under the box.
  2. Search test. Search the document for a word you know was covered (a surname, the last four digits, a company). Text that exists can be found.
  3. Extract test. Copy-all into a plain text editor, or run pdftotext file.pdf and read the output. This is the most honest of the three — it shows exactly what a machine sees.

A properly redacted file fails all three in the right way: there's nothing to select, searches come up empty, and extraction returns no trace of the covered content.

What actually needs to be removed?

Leak channelWhat it exposesDoes rasterize-and-rebuild fix it?
Visible text objectsThe words on the page, including under any cover-upYes — pages become images
Document propertiesAuthor, title, producer app, creation datesYes — output is a new document
Comments, annotations, form fieldsReviewer notes, tracked values, hidden field dataYes — none are carried over
Hidden layers and attachmentsOCG layers with originals, embedded filesYes — dropped with the page objects
Text the redactor forgot to boxWhatever it isNo — no method can; review before export

That last row is the humbler truth about redaction tools: every one of them, from Acrobat on down, destroys exactly what you tell it to. The tool's job is making destruction real; spotting every instance of the sensitive content is still yours.

How does rasterize-and-rebuild work?

There's one method that's unconditionally safe and needs nothing but a browser: re-render each page as an image, burn the redaction boxes into that image, and assemble a brand-new PDF from the results. The page image contains only pixels — no glyphs, no text instructions, nothing to select or extract. Boxes painted in before the image is encoded never contained the covered content in the first place. And because the output document is built fresh, the original's metadata, comments, and hidden layers simply don't come along.

The trade-offs are real and worth knowing. The output loses its text layer everywhere — remaining text stays readable to humans but not to search, screen readers, or copy-paste. File size usually grows, since pixels cost more than characters. Professional tools take the surgical route instead (deleting the specific content objects and scrubbing metadata) and keep the text layer for everything you didn't redact — that's what you're paying for in Acrobat. The raster approach is what a free browser-only redaction tool can do honestly, and it's the same end state as the print-and-rescan workflow people have trusted for decades, minus the printer.

Does the redaction hold up in practice?

Verification is the habit that separates a redacted file from an assumed-redacted one. After exporting, open the result and run the three tests on a couple of the boxes — select, search, extract. Check the document properties dialog too: it should show nothing meaningful about the original. If you're preparing a filing or a regulated disclosure, the applicable rules govern what counts as proper redaction, and checking them (or asking someone who knows them) comes before any tool. Nothing here is legal advice.

Not legal advice. This guide explains how the technology works. For court filings, HIPAA disclosures, FOIA productions, or contractual obligations, follow the specific rules that apply to you.

What about annotating instead?

Different job, different tool. Annotation marks things up — highlights, notes, drawings — and keeps the document's text intact, which is what you want for review and collaboration, and precisely what you don't want when the content must disappear. Our PDF annotate tool covers that case, writing standard PDF annotations in the browser. And once your redacted file is final, Organize PDF handles any reordering, page deletion, or rotation — also without uploading anything.

Redact a PDF right now

Draw boxes over sensitive text and download a rebuilt PDF where the covered content is destroyed, not hidden. Free, private, in-browser.

Redact PDF →

The bottom line

Covered is not removed. Test every redacted file with select, search, and extract before it leaves your hands; expect metadata and hidden layers to leak unless the method handles them; and when the stakes are real, prefer methods that destroy — rasterize-and-rebuild or a professional redaction tool — over anything that merely draws on top of the page.

Advertisement