Stamp text across every page, with the colour, size, opacity and position you choose. It is drawn into the page content, so it travels with the document, and it is a label rather than a lock.
Runs in your browser. Nothing is uploaded.
The text you type is drawn onto each page as ordinary page content, using an embedded standard font, at the size, colour, opacity and position you set. It becomes part of the page description rather than an annotation layered on top, so it prints, it survives being sent on, and it appears in every viewer rather than only the one that made it. The cost is small and predictable: stamping twelve pages added 2,917 bytes and forty pages added 9,373, working out at roughly 240 bytes per page for the drawn text plus the font that has to be embedded once.
Worth being clear about what this is for. A watermark is a visible label that communicates status, marks a copy as a draft, or discourages casual reuse of a document. It is not protection. The text sits in the page content stream where any PDF editor can select and delete it, and anyone can render the page to an image and work from that. If a document genuinely must not be redistributed, watermarking is the wrong mechanism and the answer is access control rather than ink.
Choose a PDF
The same questions a security review would ask, answered plainly.
Pick the exact shade, point size and transparency. A pale grey at low opacity sits behind the text without making it unreadable, which is what most documents want.
Place the stamp where it will be seen without covering the content that matters. Centre is the default; corners suit documents with dense middles.
Text is drawn onto the file you opened, not onto a replacement assembled from it, so a form still accepts input afterwards and the outline is still there.
Things people mark CONFIDENTIAL are precisely the things that should not be uploaded to mark them confidential. This does the work locally.
There is no upload step and no server to receive one. The page loads once, then all work on your PDFs happens in local JavaScript.
No daily limits, no per-file ceilings imposed by a server tier, no watermarks, and no account. The practical limit is your own device's memory.
What the stamp costs, and how that scales with page count.
The default settings on a twelve-page report.
Input
12-page text PDF, 50,563 bytes
Output
53,480 bytes, 5.8% larger, 25 ms
Settings: CONFIDENTIAL, centre, 48pt, 30% opacity, grey
2,917 bytes for twelve pages, which includes embedding the font once. The proportion looks noticeable only because the source is small; on a scan the same stamp is a rounding error.
Showing how the cost scales with page count rather than file size.
Input
40-page text PDF, 165,897 bytes
Output
175,270 bytes, 5.6% larger, 24 ms
Settings: CONFIDENTIAL, centre, 48pt, 30% opacity, grey
9,373 bytes across forty pages, about 234 a page, and no slower than the twelve page run. The per-page cost is the drawn text; the font is embedded once regardless of length.
Real files, real numbers, and the machine they were run on.
| Scenario | Input | Output | Time | Notes |
|---|---|---|---|---|
| CONFIDENTIAL across 12 text pages | 50,563 bytes | 53,480 bytes | 25 ms | 2,917 bytes added |
| CONFIDENTIAL across 40 text pages | 165,897 bytes | 175,270 bytes | 24 ms | about 234 bytes per page |
How this was measured: Measured by driving this page in headless Chrome 152 on an Apple M4 Pro running macOS 15.5, on 2026-09-17. The fixture is handed to the page's own file input, the page's own button is pressed, and the figure recorded is the size of the file the page produces. Timings come from a warm page on a fast machine with no network involved, so treat them as a floor rather than a typical result.
A document labelled DRAFT cannot be mistaken for the signed version when both are sitting in someone's downloads folder three weeks later.
Proposals and pricing sent to several parties are easier to keep track of when each is visibly marked, and the marking discourages casual forwarding.
Stamping a scan as a copy, or marking who it was issued to, is standard practice for records that could otherwise be passed off as originals.
Specimen papers and worked examples marked as samples stop them circulating as the real thing, which matters more than it sounds for assessment material.
The documents most likely to need a confidentiality stamp are the ones least appropriate to upload to a free web service. Doing it locally removes the contradiction.
Everything needed is in the tab after the first visit, so documents can be marked up on a train with no signal at all.
Starting with the one that matters most.
The text is ordinary page content. Any PDF editor can select and delete it, and rendering the page to an image sidesteps it entirely. Use it to label and to discourage, never to secure. Real restriction needs access control, not ink.
There is no option to stamp a company mark or a signature graphic. Only typed text using a standard embedded font can be drawn.
The same text, size, colour and position are applied to every page. Different wording on different pages, or skipping the cover, needs the document splitting and reassembling.
The stamp is drawn horizontally in one place per page. The diagonal repeated pattern some tools produce is not available here, which also makes it marginally easier to crop around.
No, and this page would be a strange one to get that wrong on. The file is opened in this tab, text is drawn onto each page and a stamped copy is written to your downloads. Documents being marked confidential are exactly the ones that should not pass through someone else's server on the way.
Yes, with a little effort. The text is part of the page content stream, so a PDF editor can select and delete it. Treat it as a label that communicates status and discourages casual reuse, not as a control that prevents anything.
Roughly 240 bytes per page. Twelve pages grew by 2,917 bytes and forty by 9,373, the difference being that the font is embedded once regardless of length. On a scanned document that is far below a rounding error.
Yes. It is drawn into the page rather than added as an annotation, so printers and every viewer treat it as part of the content. This is the behaviour you want; annotation-based stamps are sometimes skipped when printing.
No, text only. The watermark is typed and drawn using a standard embedded font, and there is no option to place a graphic. A logo stamp needs a full PDF editor.
Not in one pass, since the settings apply to the whole document. Splitting the file, stamping each part separately and merging the results is the route, though merging will cost you any form fields.
Around 30 percent, the default, is readable as a watermark while leaving the text underneath legible. Higher values start to interfere with reading; much lower and the mark stops being noticed, which defeats the purpose.
Not at all. The existing content is untouched and the watermark is drawn over it, so text remains selectable and searchable exactly as before, including the words sitting behind the stamp.
Yes. Stamping edits the file you opened instead of assembling a replacement, which means everything registered once for the whole document stays registered: fields, the outline, the lot. Send the same file through reorder or merge and you would lose them.
Yes, once the page has loaded. The font is one of the standard set the PDF library carries, so nothing is fetched while you work.
There is no limit imposed by us, because there is no server tier to enforce one. The real constraint is your device's available memory, since the whole file is held in RAM while it is processed. Very large files are slower on phones than on a laptop.
No. There is no sign-up, no email, and no usage tracking tied to an identity. The tool works the first time you open it.