Add page numbers

Skipping a cover takes two fields: set Start from page to 2 and leave First number at 1. Pages before the start page are left completely alone — nothing is drawn on them, and the preview leaves them clean too, so page one really does stay blank. First number controls what the first stamped page is labelled, which is why that pair produces an unnumbered cover followed by a body opening at 1. The {total} placeholder follows the same logic: it counts only the pages that actually get numbered, so a 13-page file starting on page 2 reads "1 of 12" rather than "1 of 13". If the front matter needs a sequence of its own, split it off with Split PDF, number each part separately, and join them again with Merge PDF.

Yes — Format is a small template rather than a fixed menu. Whatever you type is printed literally, and two placeholders are substituted for each page: {n} becomes that page's number and {total} becomes the count of numbered pages. Page {n} of {total} gives "Page 3 of 12", {n}/{total} gives "3/12", and Contract A — page {n} gives a prefixed label. Clearing the field draws nothing at all, which is a quick way to look through a file without stamping it. Text is set in Helvetica at the size you chose, and its width is measured before placement, so a long format string still aligns correctly in the corner you picked instead of running off the edge of the paper.

Roman numerals and letter sequences are not generated: {n} always resolves to an ordinary Arabic number, so i, ii, iii or A, B, C are out of reach in one pass. A fixed prefix or suffix is easy, though, because everything outside the placeholders is printed verbatim — A-{n} produces A-1, A-2, A-3, and Appendix {n} behaves the same way. When a document genuinely needs two sequences, such as Roman front matter followed by an Arabic body, split it with Split PDF, run each part with its own prefix and starting value, then rejoin with Merge PDF. Legal exhibit stamps with their own numbering conventions are handled by the separate Bates numbering tool rather than here.

Page numbers are drawn over whatever is already on the page — nothing in the original is moved, cleared or replaced, so an existing footer stays right where it was, underneath. That is exactly why the six position chips and the live preview matter: choose a corner that is empty, then step through a few pages and look before you download. Placement uses a fixed 28-point margin measured from the visible edge of the page as a reader sees it, so rotated and landscape pages behave properly; the tool reads each page's rotation and crop box and turns the text to stay upright rather than letting it run down the margin. Mixed page sizes in one file are fine, because every page is measured on its own.

The component reads the File object, draws the numbers with pdf-lib and creates the downloaded bytes in the browser; it does not post the selected PDF or the numbering settings to a processing endpoint. The site and any lazily loaded tool code still need normal network access unless already cached, so local document processing should not be read as a blanket offline promise.

Add page numbers is free with no account, no email step, no per-file quota and no paid tier holding anything back — the format field, all six positions and the full 6-to-96-point range work on the first run. Nothing is added to the output beyond the numbers you configured: no logo, no footer credit, no "made with" line hidden in the metadata. The saved file keeps your original page content intact and is written as an ordinary PDF that any reader opens. The only change is the text you asked for, in the corner you chose, on the pages you selected. If you ever find PDFChamp wording in a file that came out of this tool, that is a bug worth reporting.