About Word to PDF

Drop between one and 20 .docx documents, up to 25 MB each, and get a PDF for each one. Each document keeps the page size and margins it was written with, or you can force A4 or Letter for the whole batch; tick one box and the batch becomes a single combined PDF instead, in the order the files are listed.

The output is a text PDF, not a picture of one. Every paragraph is written as PDF text objects, so the result is selectable, searchable, and small: a twenty-page report usually lands under 200 KB.

A .docx file is a zip archive of XML parts. Converting one means four passes, all of them in this browser tab:

1. Unpack. The archive is inflated in memory and word/document.xml is parsed, along with the style sheet, the list definitions, the theme and the pictures the document references.
2. Resolve. Almost no formatting in Word sits on the text that shows it. A paragraph inherits from the document defaults, then the default paragraph style, then a chain of named styles, then its list level, and only then from its own properties. Walking that chain in the right order is what makes a heading come out as a heading rather than as body text.
3. Lay out. Each paragraph is measured against the real width of the font at the real page width, broken into lines, justified if it should be, and stacked until the page is full. Table rows are measured cell by cell and split across a page boundary when they have to be.
4. Write. The positioned lines become PDF text, rectangles and image objects. Latin text uses the fonts every PDF reader already has, so nothing is downloaded; Greek or Cyrillic text pulls a subsetted Noto face, which is why a Ukrainian document weighs a little more than an English one.

No server takes part in any of this. There is no API to call and no queue to wait in — the four passes above are the whole program, and they run on your processor.

Being clear about the gaps is more useful than claiming there are none. The converter reports each of these on the file it applies to, rather than letting you find out later:

• Fonts. Calibri, Cambria and their kin are licensed typefaces that cannot be redistributed by a web page, so text is set in a metrically close serif, sans-serif or monospaced face. Line breaks can therefore land a word away from where Word puts them.
• Headers, footers and footnotes. These live in separate parts of a .docx and are not drawn into the PDF.
• Text boxes, charts and SmartArt. Floating drawing canvases are skipped; ordinary inline pictures are not.
• Alphabets beyond Latin, Greek and Cyrillic. Chinese, Japanese, Korean, Arabic, Hebrew and Indic scripts need font files far larger than a web page should ask you to download, so they are not supported yet.
• Vector pictures. WMF and EMF metafiles, which older documents use for diagrams, have no PDF equivalent and are left out.

Your documents are not uploaded. The four passes described above all run on your device, and no filename, no byte of content and no page of output is sent anywhere, logged, or stored by Word to PDF. That is a consequence of the design rather than a policy: there is no server-side converter here to send anything to.

Because nothing is transmitted, the tool keeps working with the network disconnected once the page has loaded — with one exception: a document in Greek or Cyrillic needs a font file from this site the first time, so fetch that while you are still online. When you close the tab, remove a file from the list, or reload, the in-memory copies of your document and its PDF are released. Nothing is written to storage on your device, so nothing survives to the next visit.

This describes what Word to PDF does with your files, which is nothing beyond converting them for you to download. It cannot say anything about what a browser extension, a network proxy or the device itself might see — only that this tool adds nothing to that list.