Guide

How to redact a PDF

Real redaction removes the underlying bytes — not just covers them. Critical for anything going to a third party.

Every couple of years a major organisation publishes a redacted document, somebody copies the “hidden” text out of it, and the story makes the front page. The cause is almost always the same: a black rectangle was drawn over the sensitive text instead of actually removing it. The rectangle was a picture; the text underneath was still there. This guide walks through how to redact a PDF properly with Flint's Redact PDF tool — meaning the underlying bytes are removed from the file, not painted over — and the small set of habits that keep you out of that next headline.

The classic mistake: drawing a black rectangle

Open a PDF in a basic editor, draw a black rectangle over the bit you don't want anyone to see, save the file. Visually it looks redacted. But the original text is still sat in the file's content stream — the rectangle is just a shape drawn on top. Two ways anyone gets to it:

  • Copy and paste. Select the rectangle in most PDF viewers, hit copy, paste into a text editor. The original characters come out. They were never gone.
  • Move the rectangle. Open the file in any PDF editor (including free ones) and the rectangle is a movable object. Drag it aside, read the text underneath.

This is not a hypothetical. It has happened to governments, banks, law firms, intelligence services and the New York Times. The fix is to use a tool that understands the difference between “covering” and “removing”.

What real redaction actually does

A proper redaction tool does three things:

  • Removes the text from the content stream. The characters that made up the redacted word are deleted from the PDF, not hidden. Copy/paste yields nothing because nothing is there.
  • Replaces the area with a visible mark. A black bar (or a coloured one, depending on convention) shows where the redaction occurred, so the reader can see that something was removed. This is important — you want the redaction to be obvious, not hidden.
  • Strips associated metadata. A PDF may carry hidden text layers, comments, bookmarks, form field values and document metadata that reference the same sensitive content. A serious redaction pass cleans all of those.

Flint's Redact PDF tool does all three. When you mark a region for redaction, the bytes representing that text are removed before the file leaves the editor. There is no “move the rectangle aside” option because there is no rectangle in the traditional sense — it's a permanent edit.

How to redact a PDF in Flint

Open the Flint Redact PDF tool in any browser. The tool is built to be opinionated about the right behaviour — the “quick” option does the destructive thing, because the alternative is unsafe.

1

Open your PDF in Flint

Drop the file onto the upload card or drag it in from your desktop. Files up to 250 MB are supported on Pro. Once loaded, the document opens in the Flint editor with the redaction tool active.
2

Mark the regions to redact

Drag a box around every piece of text you want removed — names, account numbers, addresses, dates of birth. Mark across pages if you need to. You can also search for a specific term (a name, a phone number) and redact every occurrence in one click — handy for long documents.
3

Apply the redactions — permanently

Click Apply. Flint removes the marked content from the PDF's content stream and replaces each region with an opaque mark. The result is a new PDF where the original text no longer exists — not visually, not in the underlying bytes, not in any copy-pasteable form.

What you should be redacting

It's easy to focus on the obvious (names, addresses) and miss the trickier categories. A defensible redaction pass covers all of these:

  • Direct identifiers. Full names, email addresses, postal addresses, phone numbers, dates of birth, photographs.
  • Financial identifiers. Card numbers, bank account details, sort codes, IBANs, tax numbers, salary figures where they aren't relevant.
  • National identifiers. Passport numbers, driving licence numbers, NHS numbers, NI numbers, social security numbers — any government-issued ID.
  • Quasi-identifiers. Job title plus company plus location can identify someone as effectively as a name. Even an unusual age in a small company can be a giveaway.
  • Inadvertent identifiers. Signatures, handwritten notes in the margin, the metadata field showing who saved the file. Easy to forget.
  • Legally privileged content. Anything covered by client-attorney privilege, commercial confidentiality, or a court order to seal.

After redacting — three checks worth doing

  1. Sanity-check by copy-pasting. Open the finished PDF, try to select text in the redacted regions, paste somewhere. You should get nothing. If you get text, the redaction didn't take — open the file back in Flint and re-apply.
  2. Strip metadata. The PDF's metadata (author, original title, the application that created it) can leak who wrote the document. Make sure this is sanitised — Flint clears it as part of the export.
  3. Consider compressing. Redacted PDFs are sometimes shared widely as proof of disclosure. Run the finished file through Compress PDF if you need it to travel well.

Methods that look like redaction but aren't

Drawing a black rectangle in a PDF editor

Already covered above. Don't do this. The text underneath survives.

Highlighting text in black

Same problem. The highlight is a colour overlay; the text is still there.

Cropping the page

Cropping moves the page boundaries in but doesn't delete the content outside them. Many readers honour the crop; some can be coaxed into showing the “hidden” margins. Not safe.

Printing to PDF after applying a black box

This sometimes works, sometimes doesn't — depends on the printer driver and the source application. Some drivers flatten everything to an image, which is genuinely safe (but also makes the document unsearchable). Others preserve the text layer underneath the visual block. Don't rely on it without testing every time.

Image-only / flattened PDFs

Converting the PDF entirely to images strips the text layer, which does remove the underlying characters. But it also kills accessibility, searchability and selectable text — a heavy price for what proper redaction does gracefully.

Other tools that do real redaction

Adobe Acrobat Pro

Acrobat has a dedicated redaction tool and has had one for years. It does the right thing — actually removes text. The catch is the licence cost, and the fact that it's a desktop app you have to install.

Foxit, Nitro, PDF-XChange

The main commercial PDF tools all have proper redaction in their Pro tiers. Similar tradeoffs to Acrobat: desktop-only, paid.

Flint (the case for it)

Browser-based, no install, and the redaction is byte-level by default rather than as a paid extra. The same workspace also handles signatures, compression, splitting, merging — useful when you're cleaning a document for disclosure and need to do more than just one thing.

Tips for a defensible redaction

  • Redact a copy, never the original. Keep an unredacted master in case you need to revise. Once the redaction is applied, it's permanent.
  • Search every variant of every name. “John Smith”, “J. Smith”, “Mr Smith” and “Smith, J.” are all the same person. Use search-and-redact for each.
  • Don't forget headers, footers and watermarks. Names often appear in page headers throughout a long document. Easy to miss when you're focused on the body text.
  • Check the bookmarks and the table of contents. If a chapter title contains a name, it'll show in the bookmarks panel even after the body is redacted.
  • Treat the resulting file as the only file. Delete or quarantine intermediate drafts. The whole point of redaction is that no copy of the unredacted content travels onward.

Redacting PDFs: frequently asked questions

Is Flint's redaction reversible?

No. Once you apply, the marked content is removed from the file. That's the whole point. Keep a copy of the original somewhere safe if you might need to revise.

Does Flint redact images as well as text?

Yes. You can draw a redaction region over any part of a page — including embedded images. The region is rasterised over and the underlying image content for that area is replaced.

What about scanned PDFs without a text layer?

For scanned documents that are just images, redaction is still effective — the pixels in the redacted area are replaced with the redaction mark. The image-based text underneath can't survive, because there's nothing underneath to survive.

Will the redactions show up in print?

Yes. The redaction marks are part of the document, visible on screen and on paper. That's the intent — the reader should see that something was redacted.

Can I redact a password-protected PDF?

You'll need to unlock it first using the password you legitimately own. Once unlocked, the redaction flow is the same.

Is the file private?

Yes. The file is stored in your private Flint library, never shared, never used for training. Delete from My Documents when you're done.

What's the maximum file size?

Flint Pro accepts files up to 250 MB.

Ready to redact?

Drop your PDF into Flint's Redact PDF tool and you'll have a permanently redacted file in seconds — the kind of redaction that survives copy-paste, survives metadata inspection, and survives being opened in any other PDF tool.

Ready to try it?

The whole flow is one page. Drop your file in, get the result in seconds — no signup required to start.

More guides

How to Redact a PDF — Permanently Remove Sensitive Text | Flint — Flint PDF