Removing a Whole Page From a PDF vs Redacting It

Most redaction work is fine-grained. A name here, an account number there, a paragraph on page nine. You draw a rectangle, you check it landed, you move on. But every so often you open a page and there is nothing on it to keep. The whole sheet is one privileged memo. It is an unrelated document that got scanned into the middle of the bundle. It is somebody else's medical history that ended up in the same file as yours. The useful content on that page is zero, and the black boxes would cover all of it.

At that point the instinct is to get rid of the page. It feels tidier, it feels more thorough, and it is almost always the wrong instinct — not because deleting pages is hard, but because a document with a page missing behaves differently from a document with a page blacked out, and the difference shows up in the recipient's hands rather than in yours.

This is a decision, not a technique. It is worth separating out from the rest of the redaction process because the two options produce different artefacts, carry different obligations, and fail in different ways.

Two different things called "removing a page"

People use one phrase for two operations that have almost nothing in common.

Withholding means the page is not in the document you hand over. The file is shorter. Nothing in it says anything was taken out. From the recipient's point of view it is simply a document of that length, and whether they notice the absence depends entirely on whether something else in the file points at the gap.

Full-page redaction means the page stays exactly where it was and is covered in full. The page count is unchanged, the pagination still runs straight, and the recipient can see, unambiguously, that there is a page here and they are not being shown it. It is not a weaker version of removal; it is a statement where removal is a silence.

The second one is what this tool does, and the reason is structural rather than philosophical.

What this tool can and cannot do

Reading the export code settles it. When you press Download, the routine walks the document from page one to the last page in a plain counted loop with no condition attached, renders each page, burns in any boxes you drew on it, and adds it to the new PDF. Every page goes in. There is no page selector in the interface, no delete control, no way to export a range, and no path through the code that skips a page. The file you download always has exactly as many pages as the file you opened.

So if your plan was "drop it in here and take out page 4," that is not an operation this page offers. What it offers instead is the thing that is usually the better answer anyway: cover page 4 completely, and hand over a twelve-page document in which page 4 is a black rectangle.

If you do need a genuinely shorter file — pages extracted, deleted, reordered, split into volumes — that is a separate job for a page-management tool. The sequencing matters: redact first, then rearrange. Redacting afterwards means handling a file whose pages have already been copied, renumbered and possibly re-saved somewhere else while still carrying the sensitive content.

Why a missing page is harder to hide than it looks

The appeal of deletion is that it seems to leave no trace. In practice a page is woven into the document around it, and pulling it out leaves loose threads on every side. Before you remove a page from anything, read the remaining pages for all of these:

• Printed page numbers and "page 4 of 12" footers. These are drawn into the page content, not stored as file properties, so no amount of file-level editing updates them. A document whose footers run 1, 2, 3, 5 has announced the deletion on every sheet; a document whose footers all say "of 12" when there are eleven pages has announced it even more loudly.
• Tables of contents, indexes and exhibit lists. A front-matter entry pointing at a page the reader cannot find is a direct label for what you took out — often including its title.
• Cross-references in the body. "See the schedule at page 6," "as defined in clause 12," "continued overleaf." Contracts and reports are full of these and they do not update themselves.
• Content that runs across the break. A sentence that stops mid-clause, a table whose rows resume without a header, a numbered list that jumps. This is the most common tell and the easiest to miss, because you are looking at the page you deleted rather than the two either side of it.
• Sequential identifiers. Bates numbers, production control numbers, invoice line numbers, numbered paragraphs, form field numbering. Any counter that skips a value is a gap with a size attached to it.
• Bookmarks and internal links. In a file where these survive, destinations either break or quietly resolve to the wrong page, which is worse because nothing looks wrong.
• Arithmetic. If the deleted page carried rows that feed a total, the total still shown no longer reconciles with the rows still visible — and the difference is a measurement of what you removed.
• Attestations that count pages. Signature blocks, initialling lines, notarial certificates and transmittal letters often state a page count or require initials on each page.

None of that is detected by software. It is a reading task, and the file can be perfectly well formed while every remaining page points at the hole.

A full-page redaction sidesteps the entire list. Footers still count correctly because the page is still there — covered, but counted. Cross-references still land. Content still joins up around it. The reader knows precisely where the withheld material sat and how much of it there was, which is usually what you want them to know.

What formal processes actually ask for

Where someone has a right to the material, the rules tend to push in the same direction: say what you held back rather than quietly shortening the document.

In federal civil litigation in the United States, Rule 26(b)(5)(A) of the Federal Rules of Civil Procedure provides that a party withholding otherwise discoverable information on a claim of privilege or trial-preparation protection must "expressly make the claim" and "describe the nature of the documents, communications, or tangible things not produced or disclosed — and do so in a manner that, without revealing information itself privileged or protected, will enable other parties to assess the claim." A page that simply is not there does neither of those things.

The federal Freedom of Information Act is more specific still about the artefact. Section 552(b) of Title 5 of the U.S. Code requires that "any reasonably segregable portion of a record shall be provided" after deleting exempt portions, and that "the amount of information deleted, and the exemption under which the deletion is made, shall be indicated on the released portion of the record," unless that indication would itself harm the protected interest — adding that where it is technically feasible, the indication "shall be indicated at the place in the record where such deletion is made." That is a statutory preference for a marked blank over a vanished page.

Those two are concrete because they are checkable. Most other contexts — state courts, regulators, agency procedures, contractual disclosure obligations, internal investigations — have their own rules, and they vary. Treat the direction of travel as general and the specifics as something to confirm against the rule, order or protocol that actually governs your matter.

Choosing between them

1. Ask whether the page was ever in scope. If it is an unrelated document that was scanned into the same bundle by accident, it was never part of what you were asked to produce, and leaving it out is not withholding — it is correcting the bundle. That is the clearest case for removal.
2. If the page is responsive but entirely sensitive, redact it in full. This is the standard case and the standard answer. The page is part of the document; the content is not for the recipient; the mark records both facts.
3. Check whether anything is salvageable first. A page that looks wholly sensitive often has a usable date, heading, or document type that helps the recipient understand what they are not seeing. Covering everything including the letterhead is sometimes necessary and sometimes just coarse.
4. Add a placeholder where the convention allows it. Many processes use a short page reading something like "This page withheld in full" with a reason code. It turns a black page into an explained one, and it is the practical way to meet a requirement to describe what was held back at the place it was held back.
5. Record the decision either way. Whether you blacked a page out or left it out, write down which page, and why, before the file goes anywhere. The exported file will not tell you later.
6. Count the pages at the end. Open what you are about to send and check the page count against what you intended. This catches both the page you meant to cover and did not, and the page that disappeared in some intermediate step you had forgotten about.

Blacking out a full page here, in practice

Open the PDF, use the Prev and Next controls to reach the page, then click at one corner of the page image and drag to the opposite corner. The box you get is exactly the rectangle you dragged: the tool does not snap to the page edge, does not pad the rectangle outwards, and does not round anything up. A drag that stops a few millimetres short of the margin leaves a live strip of header, footer or page number showing — and the page number is often precisely the thing you were worried about. Drag out past the corner rather than aiming at it.

Then check the status bar. It shows a box count for the current page, which is the simplest cross-check available: every page you intended to cover in full should read one box, and a page reading zero is a page you walked past. Work through the document page by page rather than trusting memory, then download and open the result to confirm.

Common mistakes and misconceptions

Thinking a shorter file is a safer file. It is a quieter file, which is different. Removal hides the fact of the withholding, not the withholding itself, and when the gap is found the issue stops being "was this properly redacted" and becomes "what else was taken out without telling us."

Deleting the page in one tool and redacting in another, in that order. If you delete first, the pages you keep are still unredacted when they get copied into a new document, and in many page-management tools that copy includes the original text objects. Redact first, so the content is already gone from the pages before anything moves them.

Assuming the page count is a file property that can be quietly adjusted. It is not an editable label; it is the number of page objects in the document. And the page numbers your reader sees are usually ink on the page, independent of the file's own structure. Changing one does not change the other.

Covering the body of the page and leaving the furniture. Headers, footers, page numbers, running titles, department stamps and file paths sit in the margins and are easy to leave uncovered when you drag a box across what looks like "the content." On a page you are withholding in full, the furniture is part of what you are withholding.

Believing a blank page is the same as a blacked-out page. Replacing a page's contents with nothing, in a tool that edits rather than rasterises, can leave the original text objects present but unpainted — invisible on screen and perfectly extractable. A burned-in black rectangle over a rasterised page has no such layer to leave behind.

Forgetting the pages nobody asked about. Reverse sides of double-sided scans, separator sheets, fax cover pages and transmission receipts often carry names, numbers and routing details, and they tend to be skimmed past precisely because they are not "the document."

Related guides

See also, or use the redaction tool above:

• Does redacting a PDF change its file size or page count?
• How to write a redaction log for a redacted PDF

Frequently asked questions