Add a watermark to a PDF

Drawn into the page, not attached to it

There are two ways to put a watermark on a PDF and only one of them is worth having. A watermark annotation floats above the page as a separate object: some readers hide it, some printers leave it out, and anyone can delete it in seconds with free software. What this writes instead goes into the page's own content stream, in the same place as the text and the artwork — so it prints everywhere, appears in every reader, survives being flattened or converted, and cannot be clicked away. Removing it means editing the page content, which is a genuinely different level of effort.

Transparency is done with a real graphics state rather than by mixing the mark into a pale colour, which matters for what sits underneath: a forty per cent watermark over a photograph lets the photograph through, instead of laying a solid grey shape on top of it. The mark is drawn after the page content, so it is always on top — there is no way to slip it behind the text, because in a PDF whatever is drawn last wins.

Text watermarks embed their font in the file, so the result opens identically on a machine that does not have it, and that font covers Latin, Cyrillic and Greek. Characters outside it are reported rather than drawn as blanks — which is why Arabic, Bengali, Chinese and Japanese marks should be made as an image instead. An image watermark keeps its transparency, so a PNG logo with a soft edge stays soft rather than arriving in a rectangle.

A page's existing rotation is respected, so a diagonal mark runs corner to corner on a landscape page as it does on a portrait one rather than appearing sideways. Tiling repeats the mark across the whole page at the chosen angle, which is the layout that actually deters a screenshot, since there is no clean area to crop to. You can also restrict the mark to particular pages — useful for stamping DRAFT on everything except the cover.

When you want something else

A watermark is a deterrent, not a protection, and it is worth being honest about the difference. Anyone determined can cover it, crop it, or reconstruct the page underneath — it raises the effort and signals intent, which is usually all it needs to do. If a document genuinely must not be redistributed, a password controls who opens it at all, and that is a different tool for a different problem.

For stamping hundreds of files, or for a per-recipient mark that identifies who leaked a document, this page is the wrong shape of work. pdftk in.pdf stamp mark.pdf output out.pdf applies a one-page PDF as a mark over every page, and qpdf --overlay mark.pdf -- in.pdf out.pdf does the same from a script. Both take a loop and a list of names, which is how per-recipient watermarking is actually done.