Add Watermark
Stamp a text or image watermark across all PDF pages.
Drop a PDF to watermark
or click to browse
Choose File
Applying watermark...
Stamp a text or image watermark across all PDF pages.
or click to browse
Nothing is uploaded. The document is watermarked and rewritten inside this browser tab.
A watermark has one job: be unmistakable without making the document unreadable. The default of 15% is deliberately low, and it is the right starting point. Around 10% is a whisper — enough to survive a screenshot but easy to read through. Around 20 to 25% is assertive, appropriate for a draft you genuinely do not want quoted. Above 40% you are competing with the text, and people start skipping the page rather than reading it.
Print behaves differently from screen. A watermark that reads perfectly at 15% on a monitor can nearly vanish on a laser printer. If the document is going to be printed, test a page and expect to raise the opacity a little.
The −45° diagonal is conventional because it crosses the text block rather than sitting inside any one line, so it stays legible without landing on a paragraph. Horizontal (0°) is the better choice for short words in a header or footer position, and for documents in scripts where a diagonal sits awkwardly.
Size follows word length. DRAFT at 50pt spans a page nicely; STRICTLY CONFIDENTIAL — DO NOT DISTRIBUTE at the same size runs off the edge. Longer text wants a smaller size, or splitting the message so the strong word carries the weight.
Grey is the safe default and disappears sensibly on a monochrome printer. Red carries urgency and suits DRAFT and DO NOT COPY. A brand colour at low opacity looks deliberate on client-facing documents. Any colour works — the picker takes whatever you like.
The text is drawn into the page content itself, not attached as a separate annotation layer a reader could hide or toggle off. It travels with the file: forward the document, print it, screenshot it, and the watermark is there.
That makes it the right tool for marking status and provenance — signalling that a document is a draft, that a copy is confidential, that this particular version went to this particular recipient. It is a clear, visible, human-readable claim on the page, and that is exactly what most of the job needs.
It is a labelling tool rather than an access-control one. For documents that need to be genuinely restricted, Protect PDF encrypts the file with AES-256 so it cannot be opened without the password — and the two work well together: watermark for the visible claim, encryption for the actual lock. For material that must be removed from the page rather than labelled, Redact PDF deletes the underlying content.
Watermark late. Every page operation — merging, splitting, inserting, deleting, reordering — changes which pages exist, and a watermark applied before those steps ends up on some pages and not others.
The usual finishing sequence: assemble the document, add page numbers, watermark, then protect it if it needs a password. Compressing afterwards is fine — the watermark is vector text and Compress PDF only touches embedded images.
It becomes part of the page content rather than a toggleable layer, so it is not something a reader can switch off in a viewer. Keep your un-watermarked original — the tool never modifies your source file, so you always have the clean copy to produce a differently marked version from.
Diagonal at −45° is conventional because it crosses the whole text block without sitting along any one line, which keeps both the watermark and the text readable. Horizontal suits short words placed near an edge, and reads better in scripts where a diagonal looks awkward. Both are a single setting, so try each on your document.
Yes. Switch the watermark type to Image / logo, choose a PNG or JPG, and set its opacity, scale, rotation and whether it tiles across the page or sits once in the centre. To place a logo at one exact spot on a single page instead, Edit PDF and Add Image to PDF do that.
Yes — it is page content, so it prints. Do check the opacity before a large print run: printers render light greys more faintly than screens do, so a 15% watermark that reads well on a monitor may want raising to 20–25% for print.
No. The watermark is drawn as vector text in an embedded standard font, so it stays sharp at any zoom and adds almost nothing to the file. Your existing text stays selectable and searchable, and images keep their original resolution — nothing is rasterised.
One run covers every page, which is what a confidentiality or draft marking normally wants. For a subset, Split those pages out, watermark them, and Merge the document back together.
Yes. The text is drawn on top of the scanned page image, and because it is real text it stays crisp even though the page underneath is a picture.
No. Watermarking happens entirely inside this browser tab and the result goes straight to your downloads. Given that watermarked documents are usually the confidential ones, that is the point — go offline before you start and it still works.
Privacy: Watermarking is a local operation. Your PDF is never uploaded — the stamped file is built with pdf-lib inside this browser tab.