Guides / Remove text from a PDF permanently
How to remove text from a PDF so it cannot be copied back out
Updated August 30, 2026
There are two different jobs hiding behind the phrase “remove text from a PDF”. One is making the words invisible on screen. The other is making them absent from the file. Almost every tool does the first and lets you believe it did the second.
Why the text is still in there
A PDF page is a list of drawing instructions. Text is drawn by one that says, in effect, “set this font, move here, show these characters”. The characters are stored as data, separate from how they look.
Dragging a black rectangle over a paragraph appends one more instruction: draw a filled rectangle here. The instruction showing the characters is still in the list, still complete. Anything that reads the file rather than looking at it, which includes copy and paste, hands back every word under the box. Worse, in most editors that box is an annotation: a separate layer a reader can select, drag aside or delete. That is the failure mode behind every leaked-redaction story you have read.
The three levels of removal
- Covering. A drawn box, a black highlight, a white rectangle. Changes nothing about the data, however final it looks on screen.
- True redaction. The text-showing instruction is deleted from the page and a box is drawn in its place. The words are gone from the file, the rest of the page stays selectable and searchable. This is what a proper redaction tool does.
- Flattening. Each page is re-rendered as a picture. No text instructions are left anywhere, because no text is. Crude, total, and impossible to get subtly wrong.
The routes that actually work
- Acrobat Pro, if you already have it. Mark for Redaction then Apply genuinely deletes the content, and Sanitize Document strips metadata, attachments and hidden layers alongside it. If your employer pays for Acrobat Pro, stop reading and use it. It is the best tool for this and you already own it.
- macOS Preview, Ventura or later. Preview grew a Redact tool in 2022, meant to remove the text rather than cover it, unlike the markup rectangle that has been in the same toolbar for years and does not. Use Redact, save, then run the test below, because a silent failure here looks exactly like a success.
- LibreOffice Draw. Free, and it opens a PDF as editable objects. Click the text frame, delete it, export back to PDF. The words are genuinely gone. The catch: Draw rebuilds the layout on import, so fonts and spacing can shift. Compare against the original before sending it.
- Flatten from the command line. With Ghostscript installed, gs -sDEVICE=pdfimage24 -r150 -o clean.pdf original.pdf turns every page into an image. Device names vary between builds, so run gs -h and use whichever pdfimage device your copy lists. Cover the text first in any editor you like, then flatten, and the covering becomes permanent.
- Print it and scan it. Inelegant and completely effective. A sheet of paper carries no content stream. For two pages with a printer in the room, this is a respectable answer.
What does not work, no matter what the button says
- Print to PDF. The most common mistake. A PDF printer re-generates text as text, so your boxes survive and so do the words underneath them.
- Highlighting in black. In Word, in a browser, in a reader: highlighting is formatting. The characters stay, and stay selectable.
- “Flatten” in a PDF viewer. This usually means flattening form fields and annotations into the page, which can bake in a box while leaving the text beneath it. Read the wording carefully.
Prove it in 30 seconds
Never trust the appearance of the output. Test it:
- Copy and paste. Open the finished file, select all, copy, paste into a text editor. If one redacted word appears, the redaction failed.
- Search. Use the reader's find box on a name you removed. A hit means the text is still there.
- Extract. If you have poppler installed, pdftotext clean.pdf - prints everything the file will give up to a machine. That is what a determined reader would run.
Then check where text hides outside the page: document properties (title and author often carry a client name), attached files, and the annotation list. Flattening to images does not, by itself, remove an attachment.
If you want the flattening done for you
FileQuiet's PDF redaction runs the whole job in your browser tab. It suggests boxes over things shaped like account and social security numbers, you draw or accept them, and export re-renders every page as an image with the boxes burned in. It then re-opens its own output and counts the remaining text objects. If that count is not zero, no download is offered. Your document never leaves your machine, which for a document you are redacting is rather the point.
Black out a PDF in your browserHonest limits: the output is images, so the whole document stops being selectable, searchable and screen-reader readable, not just the redacted parts. If the rest must stay live, use Acrobat Pro's redaction instead.
Frequently asked
Does saving as a new PDF remove the text I covered?
No. Saving, exporting, printing to PDF and emailing a copy all preserve the page's drawing instructions, including the text under your box. Nothing about producing a new file removes content from it. Only deleting the text objects or replacing the page with a picture does that.
Can someone really recover text from under a black box?
Routinely, and without any skill. Select all and paste into a text editor is enough in most cases. If the box is an annotation, dragging it aside in a free PDF reader is enough. This has happened in filed court documents and published government reports more than once, done by people whose job was to get it right.
Is flattening to images overkill for a scanned document?
No, because two things still bite. Many scanners add an invisible OCR text layer that is fully copyable, and a box drawn over a scan is still an annotation that can be deleted to reveal the pixels underneath. Flattening resolves both at once.
What is the difference between redaction and just deleting the text?
Deleting lets the page reflow, so something is clearly missing but not where or how much. Redaction removes the words and leaves a marked block in their place, so the document keeps its shape and the reader can see that a specific passage was withheld. For legal and records work, the visible mark is usually required.