Word to PDF

A .docx is a ZIP of XML, and reading it is straightforward. Reproducing the page is not.

Word decides where every line breaks using its own algorithms, the metrics of the exact fonts installed on the machine, hyphenation rules, widow and orphan control, and a table layout engine that sizes columns from their contents. Reproducing that faithfully means running Word, or something that has reimplemented all of it. Any converter claiming a pixel-perfect result in a browser tab is doing one of those things on a server — which is to say, it is asking you to upload your document.

What comes across: all the text, in order; heading levels; bulleted and numbered lists including nesting; tables, with their cells; images; page breaks; the document title and author.

What does not: the exact page breaks and margins; bold and italic within a paragraph; fonts, colours and sizes you chose; headers, footers and page numbers; text boxes; footnotes; comments; charts and drawings saved as vector graphics.

Every conversion reports what it left out, per file. If your document is a CV, a report, a manuscript or a letter, the result will be a clean, readable PDF that says the same thing. If it is a designed brochure with a precise layout, this is the wrong tool and you should print to PDF from Word itself.

Tables are laid out in equal-width columns with the first row in bold and a hairline rule under each row. Every cell wraps rather than being cropped, so nothing is lost.

Word sizes columns from their contents, the available width and a layout algorithm this deliberately does not reimplement. A wide table will therefore look different from the original — but it will be complete and readable, which is the trade worth making.

Pictures are placed in the flow where they appeared, scaled to the text width and never enlarged. PDF stores JPEG and PNG directly, so those go in without being re-encoded.

Charts, SmartArt and drawings that Word saved as EMF, WMF or SVG are counted and reported rather than half-drawn — each would need a vector renderer, and a wrong picture is worse than a stated absence.

A document with tracked changes contains both the old text and the new. This produces the document as it would read with all changes accepted: deleted text is left out, inserted text kept, and the result says so.

Comments are not included. If you need the markup visible, accept or reject the changes in Word first and convert the result.

PDF has fourteen built-in fonts that every reader already has, and they cover WinAnsi — Latin letters, accented European characters, curly quotes, the euro and pound signs. Nothing else.

Writing Chinese, Japanese, Korean, Arabic, Hebrew, Thai, Devanagari or Cyrillic means embedding a font that contains those glyphs, and a font with usable CJK coverage is several megabytes. Downloading that on the chance a visitor needs it is not a reasonable trade for a page that mostly converts English.

So instead of producing a document full of blank boxes, unencodable characters become a question mark and the conversion tells you exactly how many were affected and which script they belonged to. If your text is in one of those scripts, this is the wrong tool and you should use a desktop converter — that is a more useful answer than a broken file.

The documents people convert to PDF are the ones that have to look official: CVs, contracts, invoices, tenancy agreements, medical letters, legal submissions, dissertations.

Every free "word to pdf" site takes an upload. The file sits on a server you know nothing about, under a retention policy you did not read, on infrastructure in a country you did not choose. For a CV that is careless; for a signed contract or a medical letter it is a disclosure.

Nothing here is transmitted. Turn your Wi-Fi off before you drop the file in — the conversion still works, and that is the only proof worth anything.