Add page numbers to a PDF

Number the pages of a finished PDF free, skipping the cover and starting wherever you need. No watermark is added and nothing is uploaded.

Drop files here (up to 20)

or paste from the clipboard

Options

Covers and title pages are counted in a document but almost never carry a number.

pt
pt

28 points is about a centimetre. Most printers cannot print closer than about 12.

Results appear here. Nothing is uploaded.

    How it works

    Helvetica is embedded once per document, then each page gets one text drawing operation appended to its content stream. The horizontal position is computed from the rendered width of that particular label rather than a fixed offset, because glyph widths differ in a proportional face — "11" is narrower than "10" — and a fixed offset drifts visibly once a document runs past nine pages. Everything happens in a Web Worker on your own machine.

    What to watch out for

    The number is painted on, not woven in

    It is added to the page's content stream, the same layer the rest of the ink lives on. Nothing in the document's structure learns about it, so an existing table of contents will not update and a reader's own page counter still counts from one regardless of what you started at.

    A page with no margin has nowhere to put it

    The number is placed at the distance you choose from the edge, and if the design already fills that space the number lands on top of it. Scanned documents are the usual offender, because the scan often includes the paper edge and leaves no clear band.

    Existing numbers are not detected or replaced

    If the document already carries numbers from whatever produced it, this adds a second set. There is no way to find and remove the first, because to the file they are simply ink like any other. Check a middle page before doing a batch.

    Limits

    Files up to about 100 MB work reliably. Beyond that, the ceiling is your device's memory rather than any rule we impose — the work happens on your own machine. Tested in Chrome 140, Firefox 143, Safari 18.

    Why this is a separate job from writing the document

    Numbers usually come from whatever produced the file — Word, LaTeX, Google Docs — and they are laid down while the document is still a document. Once it is a PDF that machinery is gone, and the pages are just pages.

    That matters most after a merge. Combine three chapters exported separately and you get one file whose numbering restarts twice, which is exactly the situation this tool exists for: number the finished article once, from the front, ignoring what each part thought it was.

    The cover page problem

    Almost every convention agrees that a title page is page one and does not say so. Print any bound report and check: the number appears from the second or third leaf onward, while the count includes everything.

    That is why there are two separate settings here rather than one. Pages to leave unnumbered decides how much front matter stays clean, and first number to use decides what the first numbered page says. Leave one page unnumbered and start at 2 and you get the ordinary academic layout; start at 1 instead and the body reads as though the cover were not there at all.

    Roman numerals for preface pages are a real convention and this tool does not do them — the honest answer is that they need a second numbering pass with a different sequence, and pretending a single setting covers it would produce worse documents than admitting it.

    Why the position is calculated rather than fixed

    A right-aligned number has to sit its own width in from the edge, and in a proportional typeface that width changes with the digits. Helvetica draws “1” narrower than “0”, so a document that passes page nine starts drifting if the offset is a constant.

    The width of each label is therefore measured at its actual size before the number is placed. The practical result is that the right-hand edge stays straight from page 1 to page 100, which is the sort of thing nobody notices until it is wrong.

    Nothing is uploaded

    The document is opened, drawn on and re-saved inside a Web Worker in your browser. Reports, dissertations, contracts and invoices are exactly the documents people number, and exactly the ones that should not be handed to a server to do it. Nothing is transmitted, and the file on your disk is untouched — you get a new one.

    Your files never leave your browser. Last updated 2026-08-02.