How to Save a Word Document as a PDF (Windows, Mac, iPad and iPhone)

The short answer

If you have the document open in desktop Word and it must look exactly the way it does on your screen: use File → Export (or File → Save a Copy) and choose PDF as the format. Word writes the PDF itself, using the same layout engine that drew the pages you are looking at, so pagination, page breaks, headers and footers all carry over. If you do not have Word, the built-in Print dialog on your device — choose “Save as PDF” or “Microsoft Print to PDF” as the printer — produces a PDF of whatever Word or the viewer is currently displaying.

If you have a .docx and no Word at all, a browser tool can convert it — the one linked at the bottom of this page does the whole job on your own machine, upload-free. But understand which of the two operations below you are choosing before you point it at a contract, a thesis, or anything with a page count you care about.

Two different operations wearing the same name

“Save as PDF” covers two genuinely different jobs, and they have different failure modes. This is the distinction most pages on this topic blur, and it is the reason people end up with a PDF that no longer matches their document.

Operation one is export. Word owns the layout engine: it has already decided where every line and page break falls, and exporting to PDF freezes those decisions into the PDF. What you see on screen is what the PDF contains. Every route that goes through real Word — desktop, Mac, mobile print sheets, Word Online's print preview — is this operation.

Operation two is re-layout. A .docx file is a package of content and formatting instructions, not a finished layout. A browser-based converter unpacks it, rebuilds the content as semantic HTML, and lets the browser's print engine paginate that HTML from scratch. The words are identical to your document's; where they sit on the page is decided all over again, by a different engine, with different fonts. Expect different pagination, different margins, and page breaks landing mid-paragraph where Word had kept them together.

The linked tool in this guide is operation two, and its own page says so: the result is not a Word-exact page layout. That makes it ideal for a memo, a handout or a draft you need to send as a PDF right now. It makes it the wrong tool for a 40-page contract with a referenced page count, or a document whose page breaks you have tuned by hand.

On Windows

Windows gives you two reliable routes: Word's own export, which preserves layout, and a system virtual printer that turns any app's output into a PDF.

Export from desktop Word

Open the document in Word, go to File → Export (older builds label this Save a Copy or Create PDF/XPS), choose PDF, and save. Before clicking OK, open the options: you can pick “Standard” or “Minimum size” publishing — Standard keeps editable text and is what you want for a document people will read or search, Minimum size shrinks embedded images and is for email attachments.

If the document has manual section breaks or carefully placed page breaks, this route is the only one guaranteed to honour them, because Word is printing the layout it already computed.

Print to PDF instead

Word is not open, or you are inside another app viewing the file? Press Ctrl+P and choose Microsoft Print to PDF as the printer — it is built into current Windows versions — then pick a filename. The PDF captures the print output of whatever application rendered the page, so if that application is Word, layout is preserved. If it is a viewer with substitute fonts, you get that viewer's layout instead.

What “ISO 19005 / PDF/A” means when Word offers it

In the PDF save options you may see a checkbox for “ISO 19005-1 compliant (PDF/A)”. PDF/A is the archival flavour of PDF: the file must embed its fonts and carry its own metadata so it renders the same way in twenty years, on machines that do not have your fonts installed. Turn it on for filings, records and anything that must remain visually reproducible. The costs: a larger file, and Word restricts a few features that PDF/A does not allow. For an everyday send-to-a-colleague PDF, leave it off.

On Mac

macOS treats PDF as a native output format, which gives you two independent routes besides Word's own export.

Export from Word for Mac

With the document open, use File → Save a Copy (some builds show File → Export) and select PDF as the file format. This is the same layout-preserving export operation as on Windows: Word for Mac has re-flowed the document with its own fonts, and the PDF freezes that result. If the document came from a Windows machine and uses fonts you do not have, your Mac substituted them at open time — the PDF is faithful to what you see, which may differ from what the original author saw. More on this below.

The system print dialog route

Every print dialog on macOS — including Word's — has a PDF pop-up menu at the bottom-left corner. Choose Save as PDF and the system converts the print output, with options for where the file goes and what metadata it carries. No Word required: this works from Preview, from a viewer, from anything that can print.

On iPad and iPhone

There is no “Save as PDF” button on the mobile home screen, but the Share and Print sheets give you two routes — one through Word's export, one through the system.

Send a copy as PDF from Word

Open the document in the Word app, tap the ··· (more) menu, and look under Share for the option to send or export a copy; you can choose PDF as the format instead of the original .docx, and then share the file to Files, Mail or wherever it needs to go.

The print-sheet route

The system alternative works from Word and from most apps that display documents: open the Share sheet or the ··· menu, choose Print, and wait for the print preview page. Pinch outwards on the preview — that gesture promotes the preview into an actual PDF — then tap Share again and choose Save to Files. The PDF comes out of the same print pipeline the document was rendered into, so it matches the on-screen layout.

Why the result can differ from your desktop file

Word on iPad and iPhone re-lays the document with the fonts the device actually has. When a document uses a font that is not installed, the system substitutes another family, and substituted metrics differ down to fractions of a point per character. Across thousands of characters, lines break differently, pages fill differently, and page breaks shift — so the pagination you tuned on a desktop is not the pagination the phone produces. This is not the mobile app failing; it is layout being computed again. For anything page-sensitive, export from the machine where the document looks right.

Without Word: the web apps

No copy of Word on the device? The two free web apps both produce PDFs, by different mechanisms.

Word for the web

If you can open the file in the free browser version of Word, use its Print command (on many builds, File → Print): you get a print preview of the document, then your browser's print dialog, where Save as PDF is one of the destinations. Because you are printing Word's own rendering, layout is faithful to what the preview showed — though fonts missing from your machine can substitute, as above.

Google Docs

Upload the .docx to Drive, open it in Google Docs, and choose File → Download → PDF document. This is a re-layout, and an important thing to know: even the .docx import is not a passthrough. Google's engine re-paginates the content with its own fonts, so the PDF reflects Docs' opinion of your document, not Word's. Headers, text boxes and complex layouts do not always survive the import cleanly. Good for getting a readable PDF out of a file you cannot open any other way; not for page-exact output.

When the document must look identical everywhere

Sometimes the requirement is not “produce a PDF” but “produce a PDF that is exactly the document, for every reader and every printer”. Three rules get you there.

Export from the machine the document was designed on, with the fonts installed. Fonts are the whole story behind shifted pages: open a Calibri-and-Cambria document on a Mac or a Chromebook that lacks them, and every substituted metric nudges the layout. Word can embed fonts in what it saves — there is an “Embed fonts in the file” option in its save settings — and a PDF/A export embeds them by construction, which is why archivists tick that box.

If the document is page-sensitive, count pages before and after export. Pagination drift shows up as a wrong count first and is cheapest to catch there.

As a last resort, flatten to an image-based PDF: produce the pages as images and assemble those into the PDF, or use the “as a picture” route your viewer may offer. Nothing can re-flow a bitmap. The price is steep and permanent — text selection stops working, the file is no longer searchable, screen readers cannot read it, and file size jumps — so treat it as a lock-the-visuals trade, not a default.

The in-browser converter: what it keeps and what it drops

The tool linked at the bottom of this page is operation two, done honestly. You drop in a .docx and a bundled library called mammoth parses it into semantic HTML — headings, lists, tables, paragraphs, images — entirely in your browser. Your document is never uploaded. The HTML is then placed into a hidden frame with a fixed print stylesheet and sent to your browser's print dialog, which is what actually produces the PDF; there is no JavaScript PDF library in the pipeline, and the download you have been promised is the one your print dialog offers (“Save as PDF”), not an automatic file.

Because the print engine is rebuilding the document, the page geometry is the print stylesheet's, not Word's: a fixed A4 sheet with 20 mm margins, applied regardless of what your Word setup said. Here is the complete list of what crosses the boundary and what does not.

• Survives: body text, headings, lists, tables, and the images themselves — embedded inline into the page.
• Does not survive: headers and footers, footnotes, comments, tracked changes, text boxes, columns, multi-section layouts, table-of-contents fields, embedded objects and macros.
• Changes shape: floating image placement is lost because images are inlined into the text flow, and merged cells and exact column widths are simplified.
• Typography is substituted: the print stylesheet imposes its own serif body font, sans-serif headings, a fixed text colour and fixed table borders, whatever you chose in Word.
• Input rules: it must be a real .docx. A legacy binary .doc — and a password-protected .docx, which is not readable in the browser — is rejected by the same magic-byte check that catches any other non-ZIP file, and a document that yields no content at all errors out.
• Size: there is no file-size cap in the code, but because every image rides along as inline data, a document full of high-resolution scans gets heavy fast. Expect the conversion to take longer and the print dialog to offer a bigger PDF.
• Bonus output: alongside the print route, the tool offers a standalone .html download of the converted content — handy when you would rather have the markup than the page.

Common problems

The PDF is missing page numbers, the running header, or the footnotes. They were lost by whichever tool built the PDF: Word's export keeps them, the browser converter cannot — headers, footers, footnotes and TOC fields are not part of the semantic HTML it rebuilds. If those elements matter, re-export from real Word or a print dialog.

Page breaks moved. The file came out of a re-layout — a browser converter, a Google Docs import, or Word on a machine without your fonts. Fix it at the source: export from the machine and app that render the document correctly, and check the page count before you send.

Fonts changed. The device that produced the PDF did not have the document's fonts and substituted families with different metrics. Install the fonts, embed them in the save, or export from the machine where the document looks right.

The file says it is not a valid document, or the converter refuses it. If the tool rejects your .docx, it is either a legacy binary .doc or a password-protected file; the check reads the file's magic bytes, so a renamed .doc cannot sneak through. Open the original in Word, save a copy as a plain, unprotected .docx, and convert that — or export the PDF from Word directly.

The PDF is huge. High-resolution images embedded a document's worth of pixels. Word's “Minimum size” publishing option on export shrinks them; the browser route inherits whatever weight the inline images carry, so lighten big images before converting.

You got an .html file when you wanted a PDF. That is the print step, not a failure: in the converter, the PDF comes from choosing “Save as PDF” inside the print dialog it opens. The .html download is the deliberate second output.

Frequently asked questions