Watermark PDF

Mark the pages of a PDF with text or a logo, at a size, colour, angle and opacity you choose.

Runs on your devicePDFWorkingNo account needed

What this tool does

Watermark PDF draws on top of the pages you select instead of rebuilding them. pdf-lib appends the mark to each page's content stream — text set in one of eight fonts every PDF reader already carries, or an embedded image — so everything already on the page is left exactly as it was underneath. Three layouts are offered: diagonal at 45 degrees, horizontal through the middle, and tiled, which repeats the word at 40 per cent of the size you set and a shallower 30-degree tilt, so a draft overlay covers the sheet without burying the words beneath it.

An image takes the place of the text the moment one sits in the queue beside the document, and the wording, font and layout settings stop applying. The logo is centred on every selected page, scaled to the share of the page width you pick, held inside 90 per cent of the page height so it cannot run off a landscape sheet, and turned to match whatever rotation the page carries. PNG transparency survives the embed, so a logo cut out against nothing does not arrive sitting in a white box. Only the first image in the queue is used, and anything ending .heic, .heif, .tif or .tiff is turned away on its extension alone, before anything tries to decode it, because the browser canvas that takes in every logo other than JPG and PNG reads neither format dependably.

The whole job runs in this tab — the file is read, stamped and written back on your own machine, and nothing is sent anywhere. The limitation that matters is what a watermark actually is. It is a mark laid over the page, not a lock on it: the stamp is a separate object in the content stream, so anyone with a PDF editor can select it and delete it, the text beneath stays selectable and copyable, and the wording of the stamp is itself extractable text. Protect PDF is the tool for a password, and Redact PDF is the one that destroys content rather than writing across it.

Three smaller limits are worth knowing before a long run. The standard fonts are encoded as WinAnsi, so accented Latin letters, curly quotes and the euro sign come through, while Greek, Cyrillic, Chinese, Japanese and emoji are swapped for question marks — an image is the only way to stamp those scripts here. Nothing measures the text against the sheet either: a long phrase at 60 pt is centred at its full width and simply runs past the edges, and there is no live preview, so a placement you like is usually the second or third attempt. Tiling has a harder edge, and this one is a genuine bug rather than a trade-off: the repeats are spaced four text-sizes apart, so once the size passes half the page width there is no room for even the first column, and the document comes back untouched while the result line still counts every page as stamped. Keep tiled text below about 290 pt on A4, less again on a narrower page, or use the diagonal layout, which always draws. Saving rewrites the document, which also breaks any digital signature it arrived with.

How to use it

  1. Add the PDF, and a logo if you want one. Several documents can share a queue; each is stamped with the same settings and comes back with -watermarked added to its name. An image in the queue is treated as the mark, not as another file to stamp.
  2. Type the text. It starts at CONFIDENTIAL. Empty text with no image in the queue is refused outright, and the wording is ignored altogether once an image is present.
  3. Choose a layout. Diagonal across the page, horizontal and centred, or tiled. Tiling spaces the repeats from the text size, so a larger size gives fewer, wider-set copies — and past half the page width, none at all.
  4. Set the size, font, colour and opacity. 6 to 400 pt in eight standard fonts, any hex colour, and opacity from 2 to 100 per cent. The default of 20 per cent red reads as a wash rather than a block.
  5. Name the pages, or size the logo. The pages field takes all, odd, even, a single number, a range such as 2-5, or a list like 1,4-6,last. Image width is a slider from 5 to 100 per cent of the page width.

Processing & privacy

Runs on your device. Your file is processed directly in this browser.

Common use cases

Circulating something that is not final

A pale DRAFT on every page keeps an interim version from being quoted as the agreed one, and not a word of the document itself has to change.

Marking internal material

A confidential mark across a board pack or a salary schedule tells anyone who prints it, or finds it printed, that it was never meant to travel.

Branding an exported document

A transparent PNG at about 30 per cent of the page width puts a logo on a quote or a report that came out of a tool with no template of its own.

Sending a sample before payment

Tiled text over a delivered draft — a translation, a set of plans, a research report — leaves the preview readable while making it awkward to pass off as the finished file.

Questions

Is the PDF uploaded?
No. pdf-lib reads the document and writes the stamped copy inside this browser tab, so the file stays on your machine. Once the library has loaded the first time, the tool carries on working with the network switched off.
Can the watermark be taken off again?
Yes, by anyone with a PDF editor. It is drawn as an extra object over the page rather than merged into it, so it can be selected and removed, and the content underneath was never altered. Treat it as a label, not a barrier.
Can I use my own font, or non-Latin text?
No. Eight of the fonts a PDF reader is required to carry are on the list, there is no box for uploading one of your own, and all eight are limited to the WinAnsi character set. Greek, Cyrillic, CJK and emoji become question marks, which the result line then repeats back at you rather than flagging, so stamp those scripts as an image instead.
Which image formats work as a logo?
JPG and PNG are embedded as they are. Other formats the browser can decode, such as WebP, GIF and BMP, are re-encoded to PNG on the way in. A HEIC photo is refused with a pointer to the HEIC Converter, which runs on the server; a TIFF is refused too, and nothing here reads one, so convert it to PNG or JPG outside OmniKitly first.
Can I place the mark in a corner, or at exact coordinates?
No. The three layouts put it diagonally, horizontally through the centre, or tiled over the sheet, and there is no dragging on a preview and no coordinate box. Add Page Numbers and Header & Footer are the tools that write to a named corner.
Can I stamp several PDFs at once?
Yes. Every PDF in the queue is processed in turn with the same settings, and a single image in the queue is used as the mark for all of them. The documents come back separately rather than merged.
What about a document that needs a password to open?
It will not be stamped. An encrypted file cannot be parsed here and is refused with a message pointing at Unlock PDF. Remove the password there, using one you already hold, then stamp the result.
Does stamping cost quality or size?
Nothing is re-rendered, so text stays selectable and images keep the data they came in with. The file grows only by the drawing instructions and any embedded logo, though re-saving does invalidate a digital signature the document already carried.

Related tools

  • Merge PDF — Combine several PDFs into one document in the order you choose.
  • Split PDF — Split one PDF into separate documents by page range.
  • Compress PDF — Reduce PDF file size while preserving the quality that matters.
  • PDF to JPG — Export every page of a PDF as a JPG image.
  • JPG to PDF — Turn one or many JPG photos into a single PDF.
  • Protect PDF — Add a password and set permissions on a PDF.