Convert Word to PDF

What the PDF carries over from Word

The conversion lays the document out the way a word processor does, and that decides what you get. Word headings that carry an outline level become the PDF's bookmark sidebar. Hyperlinks become real link annotations: a URL action for external addresses, a page destination for cross-references and contents entries. The file's title, author and keywords are written into the PDF's document information.

The export is also tagged. The PDF carries a structure tree, the marked-content flag that says the tags are there, and the document's language — which is what lets a screen reader announce a heading as a heading, and what an accessibility check looks for. A PDF made by printing to a generic PDF printer usually has none of that.

The browser fallback produces none of it. When our server is unreachable the page rebuilds the document locally, and that path writes text and images onto pages and carries the title across, and nothing more: no outline, no links, no tags, no language. It is a readable PDF and a structurally flat one, which is the thing to check in the result.

When you want something else

This tool sends your document to a server, and it is the only one on this page that does. If the file is something you would rather not send anywhere, the answer is a converter you install yourself: LibreOffice and run soffice --headless --convert-to pdf yourfile.docx, and nothing crosses the network.

And when the PDF has to match Microsoft Word exactly — a template with tight pagination, or a document someone will compare against theirs — use Word's own Save as PDF. Any converter that is not Word lays the document out with its own font metrics. It is very close, and very close is not the same thing, particularly around page breaks in a long document.