Drop your document into the zone below and every page appears with the exact number it will carry, in the exact corner, in the exact typeface. Choose the wording, from a plain 7 to Page 7 of 92. Skip the cover, start counting where you like, and put the number on the outer edge so it alternates the way a bound document needs. Change any setting and all of it updates in front of you. Nothing is uploaded.
Your files are read by JavaScript running on your own machine. There is no upload, no queue, and no copy on our server to delete later.
Drag a card to reorder it, or use the arrows under each card. The order here is the order in the output. Each card is a preview of the finished page, so what you see is what the PDF will contain.
Esc close ← → move R rotate
Starting
This is your processor doing the work, so the time it takes depends on your machine and not on how many other people are on the site.
The tool above is ready to use. The guide below is the part that decides whether the numbers look like they were typeset or like they were stuck on afterwards - where the number actually belongs on the page, why 12 mm from the edge is the number that matters, what to do about a cover page that must not be numbered, and how to number a document that already has its own numbering printed on it.
Two minutes, nothing to install, and no account needed for your first run. This PDF page numbering tool executes inside the page you are reading, so all four steps happen on your own machine.
Click the drop zone above or drag the file anywhere onto this page. Every page appears as a tile, and every tile already carries the number it will be given, drawn in the position and the type you have selected.
Pick a position: footer centered is the default and suits almost everything, footer right suits a report, and outer edge alternates left and right for a document that will be bound. Then choose the wording, from a bare number to Page 7 of 92.
A cover page almost never carries one. Set Which pages to a range such as 2- and the cover is skipped, while the counting still starts at 1 on the following sheet. Odd and even are there for documents printed double sided.
Click any tile to open that page at full size and confirm the number sits where you want it against the real content. Then press Add numbers. The text is written as real text, so it stays selectable and searchable in the finished PDF.
Everything in the rail is live. Change the type size and forty tiles resize at once; switch from footer centered to outer edge and the numbers jump sides on alternate pages while you watch. Nothing about the preview is an approximation: the position, the plate and the string are computed by the same code that writes the file.
There are eight positions in the rail and only four of them are ever the right answer for a given document. The choice is a typographic convention and not a matter of taste, and getting it right is most of what separates a document that looks produced from one that looks processed.
| Position | Use it for | Why |
|---|---|---|
| Footer, centered | Contracts, reports, manuscripts, anything read on a screen or printed single sided. | The safest choice in existence. It is symmetrical, so it never collides with a binding edge, and it is where every reader instinctively looks. |
| Footer, right | Business reports, proposals, decks printed as documents. | The eye finishes each page on the right, so the number is where the eye already is. It is also the convention in most corporate templates. |
| Outer edge, alternating | Anything that will be printed double sided and bound: booklets, manuals, theses, books. | The number lands on the right of odd pages and the left of even ones, so it is always on the outside of the spread and never disappears into the gutter. |
| Header, centered or right | Legal filings, technical specifications, documents with a running head already. | Some conventions and some courts require it. Otherwise a header number competes with the document title for the same attention. |
| Footer or header, left | Rarely correct on its own. | A left aligned number reads as the start of a line and gets mistaken for a list. It exists for documents that pair it with something on the right. |
The distance from the edge defaults to 12 mm, and that number is not arbitrary. Most desktop and office printers cannot print within about 5 mm of the paper edge at all, and many of them trim slightly differently from the digital page anyway. Put a number at 5 mm and it gets clipped on a printer you do not own. You will never hear about it.
Print one page before you run the whole document. Stamp two pages, print one of them on the printer that will actually be used, and hold the sheet up against what is on your screen. Thirty seconds there is what stops you discovering the clipping problem halfway through a two hundred page run.
A page number added by a good tool is real text drawn into the existing page. A bad one glues on a picture of a number, or worse, re-renders the entire page into a bitmap with the number baked into the pixels alongside everything else. You cannot see the difference. It shows up the first time somebody searches the document, copies a paragraph out of it, or feeds it to anything that reads text.
Two details make the difference between a tool that works on the document in front of you and one that works on a test file. Both are handled here, and both are where other tools quietly fail.
The first is page rotation. A page stored sideways with a rotation flag is displayed upright, and a number placed by naive arithmetic lands on what the file thinks is the bottom rather than what you can see. Here the position is computed against the page as displayed and then mapped back into the file coordinate system, with the number itself turned to match, so a rotated page gets its number in the corner you are pointing at.
The second is mixed page sizes. A report with an A3 foldout in the middle has two page sizes, and a number positioned by absolute coordinates would sit 40 mm from the edge on one and 12 mm on the other. Every position here is computed as a proportion of each individual page, so the number is 12 mm from the edge of whatever page it lands on.
The fourteen standard PDF typefaces cover Western European Latin and nothing past it: English, French, German, Spanish, Italian and the Nordic languages are fine, while Polish, Czech, Turkish, Greek and Cyrillic are not. Put a character outside that set into your custom wording and a real font file is fetched, then a subset of it is embedded into your document carrying only the glyphs you actually used. You do nothing. The only trace is a file a few kilobytes larger.
The mechanics of stamping a number are trivial. The part people actually get stuck on is counting, and the two controls that handle it are Which pages and Start counting at. They do different things and the difference is the whole answer.
2- and the first sheet gets nothing.That combination covers every normal case:
| What you want | Which pages | Start counting at |
|---|---|---|
| Number everything from 1 | Every page | 1 |
| Skip the cover, second sheet becomes page 1 | Range: 2- | 1 |
| Skip the cover, second sheet stays page 2 | Range: 2- | 2 |
| Skip a cover and a contents page | Range: 3- | 1 |
| This is volume two, continuing from page 240 | Every page | 241 |
| Number only the appendix, from A1 onward | Range: 88- | Use custom wording with {n} |
When the wording contains a total, the total is the count of numbered pages, not the count of sheets in the file. A 93 sheet document with the cover skipped reads "Page 1 of 92", not "Page 1 of 93". That is correct: the reader is being told how many numbered pages there are, and a document whose last page says 92 while the header claims 93 is a document that looks wrong to everybody who checks.
Legal discovery uses Bates numbering: a prefix plus a zero padded sequence, unique across an entire production and frequently across several separate documents inside it. The prefix and the sequence both work here on one document through the custom wording field, so SMITH-{n} produces SMITH-1 through SMITH-940. Two things it will not do: zero pad to a fixed width, and carry a sequence across separate files. That gap is deliberate. If you are producing documents for discovery, use software built for discovery. A numbering error there is a legal problem and not an aesthetic one, and a browser tool running in a tab is the wrong instrument for that job.
This is the second most common problem after covers, and the honest answer is short: a number that is already printed on the page stays there. This tool draws a new number on top of the existing content. It cannot find, move or remove a number that was typeset into the document by Word, InDesign or whatever produced it, because by the time it is a PDF that number is just marks in a content stream indistinguishable from any other text.
You have three practical options, in order of how good the result is:
Check a page in the middle, not the first one. Documents very often carry no number on page one and a number on every page after it, so a cover that looks clean tells you nothing at all. Look at page 5, and at page 40. Otherwise you discover the collision on page two of the printed copy.
Five treatments are offered and four of them are wrong for a serious document. That is not a criticism of the four: a workshop handout, a school newsletter and a product brochure are all improved by something with a bit of personality. It is a criticism of using them on a contract.
| Treatment | What it is | Where it works |
|---|---|---|
| Plain | Just the number, in the type and colour you set. | Contracts, reports, manuscripts, legal documents, anything that has to look serious. This is the default for a reason. |
| Pill | The number on a soft rounded block of colour. | Brochures, decks exported as PDFs, anything with a design language already. |
| Outline | The number inside a thin box. | Technical documentation, drawing sets, anything with a title block convention. |
| Circle | The number in a filled disc. | Handouts, worksheets, instructions where the number is a landmark rather than a reference. |
| Rule | A hairline above the number, spanning its width. | Editorial layouts and long form documents. It reads as a footer rather than as a stamp. |
The two settings people get wrong most often, in both directions:
The opacity control exists mostly for the treatments. Dropping the pill or the circle to around 12 percent gives you a hint of a plate rather than a solid block, which is what makes it look intentional at small sizes. Dropping the number itself below full strength is almost always a mistake.
The usual architecture for a free tool that adds page numbers to a PDF is an upload form, a queue, a command line utility on a server you do not control, and a temporary download link. Your file made a round trip across the internet, was written to a hard disk on the way, and was kept there for however long a privacy page claims. Look at what people actually number: contracts, tenders, board packs, thesis chapters and evidence bundles. Every one of those is confidential by default. That is precisely the category where an unplanned copy is hardest to justify to whoever asks about it later.
There is nothing to upload to. Your file is opened with the browser File API, parsed and rewritten by a PDF library running as JavaScript on your own processor, and the numbered document arrives as a Blob your browser downloads from itself, which is why the whole operation finishes with the network sitting idle. The file exists on your disk and in your RAM, and nowhere else.
The one exception, stated plainly because it is the only network request that ever depends on your content: if your custom wording contains a character outside Western European Latin, a font file is fetched from this domain so it can be embedded into your document. That request carries the name of a font and nothing else. Your document is still never transmitted.
| Symptom | Cause | Fix |
|---|---|---|
| The number sits on top of the content | The document has no bottom margin, which is common in scans that were cropped tightly and in slides exported as PDFs. | Move the number to the header, or increase the distance from the edge, or crop the page first so there is somewhere to put it. |
| The number is cut off in print | It is closer to the edge than the printer can reach. Almost every office printer has an unprintable border of about 5 mm. | Raise the distance from the edge to at least 10 mm and print one test page. |
| Two numbers on the page | The document already had its own numbering printed into it. Nothing can find and remove that. | Renumber in the source document, or place yours somewhere the old ones are not, or cover the old ones with Edit PDF first. |
| The total is one less than the page count | You skipped the cover, so it is not a numbered page and does not count toward the total. | This is correct behavior. If you want it counted, number every page and start counting at 2. |
| Characters came out as question marks | The wording contained a character outside Western European Latin and the font file could not be fetched, usually because a blocker refused it. | Disable the blocker for this domain and run it again. The font request carries no document data. |
| The card says Locked | The document is encrypted with a password. | Open it in a reader with the password, save an unencrypted copy, and number that. |
| A signature warning appears afterwards | The document was digitally signed. Any modification invalidates a signature. | Number before signing. There is no way around this in any tool, and a tool that claims otherwise is not preserving the signature. |
Numbering is usually the last step before a document goes out, or the first step in making a pile of scans into a document. Either way, these are the tools that come next:
Drop the file into the zone above. Every page appears with the number it will be given, already in position. Choose the corner, the wording and the type size, decide which pages get a number, and press Add numbers. It is free and there is no watermark on the result. No account is needed for your first run: an anonymous visitor gets one operation per 12 hours, a free CartHacker account gets two, and Premium removes the counter entirely.
Yes, and it is two controls doing two different jobs. Set Which pages to a range of 2- so the cover receives nothing, and leave Start counting at on 1. The second sheet then prints as page 1. If you want that second sheet to print as page 2 instead, set Start counting at to 2 and the cover still receives no number of its own. Which pages decides what gets stamped; Start counting at decides what it says.
Real text. A short content stream is appended to each page that sets a font and draws the string, so the number is selectable, searchable and copies out with the rest of the text. Nothing is rasterized and no page is rebuilt, which is why a hundred numbered pages add only a few kilobytes to the file and why bookmarks, links and form fields all survive.
Footer centered is the safest choice for almost any document and is the default. Footer right suits business reports and matches most corporate templates. Outer edge, which alternates right on odd pages and left on even ones, is the correct choice for anything printed double sided and bound, because it keeps the number out of the gutter. Header positions exist for legal filings and technical specifications that require them.
The default is 12 mm. That is a good answer for anything that will be printed, because most office printers cannot print within about 5 mm of the paper edge at all and anything tighter risks being clipped on a printer you do not own. Use 18 to 25 mm for documents that will be professionally trimmed, hole punched or bound. Only go below 10 mm for documents that will only ever be read on a screen.
No, and no browser tool can. Once a document is a PDF, a number typeset by Word or InDesign is just marks in a content stream, indistinguishable from any other text on the page. The good options are to renumber in the source document, to place your new numbers somewhere the old ones are not, or to cover the old ones with a white box using Edit PDF before you stamp new ones.
No. The total is the count of pages that receive a number, not the count of sheets in the file. A 93 sheet document with the cover skipped reads Page 1 of 92, which is correct, because the last numbered page will say 92 and a total that disagreed with it would look like an error to every reader who checked.
Yes. Choose Something else under Wording and type whatever you want. {n} is replaced by the page number and {t} by the highest number printed, so SMITH-{n} gives SMITH-1 onward and Contract 4471, page {n} of {t} gives exactly that. Any character your keyboard produces works, including Polish, Greek and Cyrillic, because a real font is embedded when one is needed.
Partly, and the honest answer matters here. A prefix plus a running sequence on one document works: type SMITH-{n} and you get SMITH-1 through to the end. What this tool does not do is zero pad to a fixed width, or continue a sequence automatically across separate files, both of which real Bates production requires. If you are producing documents for discovery, use software built for it. A numbering error there has legal consequences rather than aesthetic ones.
No. The numbering runs in your browser on your own processor and your document is never transmitted anywhere. Load this page, disconnect from the internet entirely, and it still works. The only request that ever depends on your content is a font file fetch, and only if your custom wording contains a character outside Western European Latin. That request carries the name of a font and nothing else.
That means the document has no bottom margin, which is normal in tightly cropped scans and in slide decks exported as PDFs. Three fixes, in order: move the number to the header, increase the distance from the edge, or crop the page first so there is somewhere to put it. You will see it coming. Every page tile above shows exactly where the number will land, so the collision is visible before you spend an operation on it.
Yes, and both are handled deliberately. On a rotated page the position is computed against the page as you see it and then mapped back into the file coordinate system, with the number turned to match, so it lands in the corner you pointed at rather than in the corner the file thinks is the bottom. On a document with mixed page sizes every position is a proportion of each individual page, so an A3 foldout gets its number measured 12 mm from its own edge, not from where an A4 edge would have been.
All of these run the same way: your file is read by JavaScript on your own machine, the work happens on your processor, and the result is handed back to you as a download that never existed anywhere else.
Find the apps that slow your store down, by name and in milliseconds, then answer shoppers' doubts with trust badges under Add to cart. Both are free, with no card and no access to your store admin.
Paste any page and get its three Core Web Vitals, the apps costing time, like Klaviyo or Judge.me, and the Shopify or WooCommerce fix.
Show the promises that settle doubts, from free shipping to a money-back guarantee, right under the Add to cart button. One line in your theme.
Change the link after printing and count the scans. 5 tracked codes on a free account, static codes unlimited.
Anonymous visitors get one operation per 12 hours, per network address. That is not a sales tactic, it is the only lever we have: this tool runs on your processor, not ours, so the run itself earns us nothing. An account tells us who you are, and the ceiling goes up the moment you have one. Your allowance resets in 12 h.
No credit card, no trial period, no sales call. Join with Google, GitHub, Facebook or email.
Free accounts get two operations per 12 hours, and yours reset in 12 h. Waiting costs you nothing and is a completely reasonable choice. If PDF work is a weekly job rather than a one-off, the counter is the wrong tool for you and Premium takes it off the page.
Cancel anytime. Everything you have already produced stays yours either way.