← All posts

Why doesn’t my converted PDF have page numbers?

You converted the document and the words all arrived. Headings are still headings, the paragraphs are in the right order, nothing is obviously broken. But the bottom of every page is empty.

Which is strange, because the whole time you were writing it, the foot of the page said "Page 3 of 12". You could see it. You may even have spent five minutes getting it centred.

It is not that the conversion lost the numbers. In almost every case they were never in the document in the first place — and once you know where a word processor actually keeps them, both the disappearance and the fix take about a minute.

A page number is an instruction, not text

Type "Page 3 of 12" into the footer of a Word document and look at what gets saved. It is not those words. It is a field: a short instruction that says print the number of the page you are currently on, sitting next to another one that says print how many pages there are altogether.

Neither has an answer until something works out where the pages fall. That is the job of a layout engine, and it only knows the answer while it is laying the document out — it flows your text, discovers that the third paragraph on page 6 spills onto page 7, and stamps the running numbers as it goes. Change the font size and every one of them changes with it. The document does not store twelve numbers. It stores one instruction, twelve times over, and Word recalculates the lot every time you touch anything.

So a converter that is not a layout engine has nothing to convert. There is no "3" anywhere in the file for it to carry across.

And the instruction lives somewhere the converter may never look

This is the second half of it, and it explains a specific symptom.

Inside a .docx — which is a zip, and you can open it as one — your prose lives in a part called word/document.xml. Headers and footers are not in there. Each one is its own separate part, header1.xml or footer1.xml, attached to a section of the document by reference rather than sitting inline with the text.

A converter that walks the document body reads every paragraph you wrote and never opens the footer parts at all. So here is the diagnostic worth ten seconds:

Either way, retrying with a different converter is the wrong move. The number is not in there to be recovered.

If your source was Markdown, HTML or plain text, there were no pages at all

A different situation with the same ending. Markdown has no concept of a page — no page break, no page size, nothing to count. HTML does not have pages either; a browser invents them at print time. The only way to put a number on an HTML page is a paged-media stylesheet with @page rules and a CSS page counter, and that is a feature of specialist print engines, not of a text file.

Our Word to PDF converter flows plain text, Markdown or HTML onto real A4 or Letter pages with real margins, and it adds nothing to those margins. No page number, no running header, no contents page. That is not an omission we are apologising for so much as a fact about the input: there is nothing in a .md file that says what the footer should contain. (Converting a README is its own small art, for related reasons.)

If you printed the thing from a browser instead and it came out with "1/12" and a URL along the bottom, those are the browser's, added by the print dialogue, not the document's. They cannot be styled, moved or renumbered. Untick "Headers and footers" and they go.

Put the numbers on the PDF, where they will stay put

Numbering the finished PDF is the reliable direction of travel, because a PDF genuinely does have pages and they are genuinely fixed. Nothing has to be recalculated ever again.

  1. Merge first, if there are several files. Number a chapter on its own and it starts at 1; number it again after Merge PDF has put four chapters together and it starts wherever it actually starts. Doing it the other way round means doing it twice.
  2. Run Add page numbers to PDF. Pick a corner — bottom centre for a report, bottom right for anything bound on the left. Pick the wording: just the figure, Page 1, 1 of N or Page 1 of N.
  3. Set the range and the start value separately. 2- leaves the cover clean; start at decides what figure the first stamped sheet gets. They are two different decisions and getting them confused is the classic mistake.
  4. Open one busy page before you file it. The number is drawn on top of whatever is already there. A page with a long footnote may need a bigger margin.

If you want the running title back too

The page number was probably not the only thing in that footer. Add header and footer to PDF puts one line along the top of every page, one along the bottom, or both, with tokens filled in page by page: {page}, {pages}, {filename} and {date}. A footer of Annual report {date} — {page} of {pages} gets you the whole line your document used to have.

Use one tool or the other in a given corner, not both. Two centred footers land on top of each other.

What these tools will not do

Convert your .docx. Our converter refuses Word files outright rather than half-rendering them, because doing it properly means running LibreOffice and it is not installed here. If that is where you are stuck, the refusal has its own explanation and the answer is below.

Build you a contents page. Create PDF bookmarks can construct a navigable outline from the finished PDF's headings, which is what most people actually want, but it is a side panel in the reader — not a printed table of contents with page numbers on it.

Change what the reader's page box shows. The number you stamp is ink on the paper. The "3 of 12" in Acrobat's toolbar comes from somewhere else entirely, and stamping does not touch it. If you need to know what your file currently claims about itself, our PDF inspector reads the page count and page sizes out of the file in your browser, without uploading it.

Come off again easily. Stamped numbers are drawn into the page content, exactly like the rest of the text. That is what makes them durable, and it is also why removing them later is an editing job rather than an undo.

When a desktop tool is genuinely better

If the numbers matter more than avoiding an install, these three do the job properly and all are free:

The short version

Your word processor was showing you a calculation, not a thing in the file. Nothing survived because nothing was there. Convert the words, then number the pages once they exist and have stopped moving.

Tools this post uses: