Someone draws a black rectangle over a client's bank details, saves the PDF, and emails it out. The page looks clean. Anyone who receives it can select the text under that rectangle, press Ctrl+C, and paste the account number into a text editor.
This is the single most common document-security mistake, and it has leaked court filings, government reports, and corporate contracts. It keeps happening because almost every tool that draws a black box calls the feature "redact" — and the result looks identical to real redaction on screen.
This guide covers what is actually happening, how to check whether a file you have received is genuinely redacted, and how to redact properly.
Why does a black box not remove the text?
A PDF page is a set of drawing instructions, not a picture. A black rectangle is one more instruction, drawn on top of the earlier ones. It hides the text from your eyes the way a sticker hides words on a page — but the text instruction is still in the file, in its original position, and any program that reads the file rather than looking at it can still read the words.
That includes:
- Copy and paste. Select across the black bar and paste elsewhere.
- Text extraction. Any "extract text from PDF" feature returns the hidden words.
- Search. Ctrl+F inside the viewer will often still match text under the box.
- Converting the file. Turn it into a Word document or an HTML page and the covered text reappears as ordinary content.
The rectangle is decoration. The data never left.
How to tell if a PDF is really redacted
Three checks, in order of speed. Any one of them failing means the file is not safe to send.
- Drag-select across the redacted area, then paste into a plain-text editor. If words appear, the redaction is cosmetic.
- Run the file through a text extraction step — our Extract Text from PDF tool returns the raw text layer of every page. Read what comes out. This catches text that sits under an image rather than under a rectangle, which drag-select can miss.
- Check the document properties. Author names, original filenames, revision software and sometimes earlier titles live in the file's metadata, entirely separate from the page. Our PDF Metadata Editor shows those fields and lets you clear them.
If you regularly receive documents that are supposed to be redacted — from a supplier, a law firm, a public body — running check 1 takes about ten seconds and is worth doing every time.
What real redaction has to do
Genuine redaction means the covered content is no longer present in the file at all. There are two established ways to achieve that, and they behave differently.
Remove the underlying instructions. The tool parses the page, works out which text and image instructions fall inside the marked area, deletes those instructions, and writes a new file. This preserves selectable text everywhere else on the page, which is good for accessibility and for search. It is also fiddly: text is often drawn in runs that straddle the boundary of the marked area, so a run has to be split rather than dropped, and a tool that gets this slightly wrong can leave fragments behind.
Convert the page to an image. The page is rendered as it appears — with the black boxes already painted on — and that flat image replaces the page content. There is no text layer left to recover, because there is no text layer at all. This is the more conservative of the two: nothing can hide behind an instruction that no longer exists.
How PDFGrover's Redact PDF works
Our Redact PDF tool takes the second approach. You draw rectangles over the areas you want gone, and every page you have marked is rasterized — converted to a flat image with the redaction already burned in — before the new PDF is written. The words underneath are not covered up; they are not in the output file.
Two things follow from that, and you should know both before you use it:
- Redacted pages stop being searchable and selectable. Because the page becomes an image, text on that page can no longer be copied or found with Ctrl+F. Pages you did not mark are untouched and keep their text.
- Redacted pages are usually larger. An image of a page generally takes more space than the instructions that drew it. If size matters for what you are sending, run the result through Compress PDF afterwards.
The whole process runs in your browser. The file is not uploaded, which for a document you are redacting is rather the point — it would be an odd privacy tool that required you to hand the unredacted original to a server first.
Redaction is not the same as cropping, flattening, or a password
These four get confused constantly, and three of them will not protect anything.
| What you do | What it actually does | Safe for secrets? |
|---|---|---|
| Redact | Removes the content from the file | Yes |
| Crop | Hides page edges by changing which rectangle the viewer displays | No — trimmed content stays in the file and can be revealed |
| Flatten | Merges form fields and annotations into the page | No — flattened field values remain readable |
| Password protect | Restricts who can open the file | No — anyone who can open it sees everything |
Cropping in particular is often used to "cut off" a header or a margin stamp. It does not remove anything; it changes a display setting. If your reason for cropping is confidentiality, redact instead.
Flattening is covered in more depth in does flattening a PDF remove signatures. The short version: it makes fields non-editable, not invisible.
A practical redaction checklist
For anything going outside your organisation:
- Work from the original, not from a copy someone has already marked up. Existing annotations may carry text of their own.
- Mark every instance, not just the obvious ones. A name usually appears in a header, a footer, a signature block and the body.
- Redact, then verify. Re-open the output and run the drag-select and text-extraction checks above on the finished file. Verify the file you are actually sending.
- Clear the metadata. Redaction works on pages. Author, subject and original filename live elsewhere.
- Check attachments and layers. Some PDFs carry embedded files or alternate layers that no page-level operation touches.
- Send the new file. The most common failure at this step is attaching the original by mistake, because it has the friendlier filename.
Common questions
Can redacted text be recovered from a PDF? If the redaction only drew a shape over the content, yes — usually by selecting and copying it. If the underlying instructions were deleted or the page was converted to an image, the content is not in the file and there is nothing to recover.
Is drawing a black box in a word processor safe if I export to PDF? Generally not. If the text is still a text object underneath the shape, exporting carries both into the PDF. Delete the words in the source document rather than covering them.
Does redacting increase file size? Often, yes, when the method converts pages to images. Compressing the finished file afterwards usually brings it back down.
Is a highlighter or white rectangle any different from a black one? No. The colour of the shape has nothing to do with whether the text below it still exists. A white box over white-background text is arguably worse, because the result does not look redacted at all.
The one-line version
If you can still select it, it is still there. Cover-ups are not redaction — removal is. Draw your boxes in Redact PDF, then verify the output before it leaves your machine.