A word printed faintly across a page has a simple job: tell whoever holds the document what it is. DRAFT says the numbers may change, CONFIDENTIAL says who should be reading, SAMPLE says this is not the paid copy. On paper the mark is ink in the fibres. In a PDF it is an object in a file, and the PDF format offers more than one place to put it. Where it goes decides whether it prints, whether a screen reader announces it on every page, whether it disappears behind a scanned image, and how much effort it takes to remove.
Three Places a PDF Can Keep a Watermark
The most common place is the page's own content: the same list of drawing instructions that puts the text and pictures on the page. A watermark added this way is drawn like anything else, so every reader and every printer shows it. Most watermarking tools, this site's included, work like this.
The second is an annotation. Since PDF 1.6 the format has a dedicated watermark annotation, meant for graphics that appear at a fixed size and position on the page regardless of the dimensions of the printed page. Its FixedPrint dictionary holds a matrix that keeps the mark the same size when the page is scaled to a different sheet of paper. Annotations sit in a layer above the page content, and readers handle them separately, so support for this type is less certain than for plain content.
The third is the one Adobe Acrobat uses for its own editable watermarks, headers, footers and stamps. Acrobat draws them as page content, but wraps each in a form XObject carrying a private PieceInfo entry named ADBE_CompoundType. That entry is how Acrobat finds its own marks again later, which is what lets it update or remove a watermark it created. To any other reader, the mark is ordinary content.
The Three Approaches Side by Side
How each way of storing a watermark behaves. Annotation details from the PDF 1.6 watermark annotation; the Acrobat entry as documented in PDFlib's PieceInfo reference.
| Stored as | Shown by every reader | Size on a different paper | Easy to find again |
|---|---|---|---|
| Page content | Yes | Scales with the page | Only if marked |
| Watermark annotation (PDF 1.6) | Depends on the reader | Fixed by FixedPrint | Yes, by its type |
| Acrobat form with ADBE_CompoundType | Yes | Scales with the page | Yes, in Acrobat |
On Top or Behind: Why Some Watermarks Vanish
A page's content is painted in order, and anything painted later covers what came before. A watermark drawn at the end of the list lands on top of everything. Drawn at the start, it lands underneath, and the page's text is painted over it.
Behind is the nicer result on a typed document. The letters stay fully black, and the mark only shows in the white space between them, much like a watermark in real paper. The preview in Watermark PDF imitates this by multiplying the mark into the rendered page, the same thing that happens when dark ink is printed over a light tint.
It fails completely on a scan. A scanned page is one picture covering the whole sheet, and its white paper is made of opaque white pixels, not empty space. Anything painted behind that picture is covered, so the saved file looks unmarked even though the watermark is in it. The same happens on pages with a filled background, such as slides exported to PDF. For those, the mark has to go on top, faded enough to read through.
How a PDF Fades a Mark
A faint watermark is not a pale colour. It is a colour painted with a constant alpha, a transparency value set in the page's graphics state, a feature PDF gained with version 1.4. The mark at 30 percent opacity lets 70 percent of whatever is underneath show through, so it looks lighter over white paper and darker over a photo.
That difference matters for archives. PDF/A-1, the first version of the ISO 19005 archiving standard, forbids transparency altogether, so a faded mark breaks PDF/A-1 compliance. PDF/A-2 and later allow it. Separately, every PDF/A version requires fonts to be embedded, and a text mark drawn in one of the built-in standard fonts is referenced by name, not embedded. If a document has to stay PDF/A, add the watermark in the program that made it, then export.
Keeping screen readers quiet
To a screen reader, a word drawn across the page is just more text, so an unmarked CONFIDENTIAL can be read aloud on every page, often in the middle of a sentence. The PDF format has a fix: content can be marked as an artifact, something that is not part of the document's real content. Since PDF 1.7 an artifact of type Pagination can also carry the subtype Watermark. Watermark PDF wraps every mark in that tag, so in a tagged document, assistive technology and text extraction can skip it.
Why a Watermark Should Barely Change the File Size
A PDF can define a piece of drawing once, as a form XObject, and place it on any number of pages by reference. Watermark PDF does that. The text or logo is stored once, and each page gets only a short instruction saying where to place it, at what angle and scale.
In a test on a 300-page document, one CONFIDENTIAL across each page added about 70 KB to the file, roughly 230 bytes per page. Tiling the same word across every page, with dozens of copies on each, added about 220 KB, still well under a kilobyte per page. A tool that pastes a fresh copy of a logo onto each page instead can turn a 300-page file into 300 copies of the image.
Watermark your PDF
Text or logo, once or tiled, on top of the page or behind it, with a preview of every page before you save.
What a Watermark Does Not Stop
A watermark is a label, not a lock. Wherever it is stored, it is a separate object on the page, and PDF editors can select it and delete it. Developers have posted scripts that find Acrobat-made watermarks by their ADBE_CompoundType entry and drop them in bulk. A tiled mark takes more work to clean up than a single one, but not much more for someone with the right software.
What makes a mark genuinely hard to remove is turning each page into a picture, so the mark and the text become the same pixels. That is sometimes called flattening, and you can do it with this site's tools by converting the watermarked PDF to images and back to a PDF. The cost is real: the text can no longer be selected, searched or read by a screen reader, and the file is usually much larger. Even then the mark can be painted out by hand, one page at a time.
So choose a watermark for what it does well. It makes the status of a document obvious at a glance, it travels with every copy and every printout, and it makes someone who wants to pass off a draft as final do deliberate work to hide it.
Before You Watermark a PDF
A few checks avoid the usual ways a watermarked file goes wrong.
Check a dense page, not the cover
A mark that looks fine on a title page can make a page of small print hard to read. Step through the preview to the busiest page and set the opacity there.
Look for scanned pages
If any page is a scan, a mark placed behind the content will not show on it. Put it on top, or watermark the scanned pages separately.
Watermark before signing
A digital signature covers the exact bytes of the file, and adding a watermark changes them. Sign the watermarked version, not the other way round.
Keep the clean original
The watermarked file is a copy for sharing. Keep the unmarked file, so a final version or a different label never has to start from a stamped copy.
Frequently Asked Questions
Related Tools
Keep Reading
PDF Page Numbers: The One on the Page and the One in the Page Box
A footer number is text on the page, and readers never look at it. Page labels decide what the page box shows, and Rotate decides where a number should go.
Turning a PDF Page Into an Image: DPI, Size and Format
A PDF page has inches, not pixels, so the DPI you choose sets the image size. What each resolution is for, why JPG smudges text, and why Word gets the size wrong.
Photos in a PDF: What Happens to a JPG, a PNG and a HEIC
A JPG can be copied into a PDF byte for byte, a PNG is compressed again and a HEIC is refused. Plus why phone photos land sideways and what sets the page size.
Merging PDFs: What Comes Through and What Quietly Disappears
A merge copies pages, not documents. Text and images arrive untouched; bookmarks, form fields and signatures do not, and the reason is in the file format.