How to convert Word to PDF without losing formatting

The reason your formatting breaks

Open almost any browser-based Word-to-PDF converter and it will ask you to choose a font, a size, and a margin before it starts. That question is the tell. A .docx file already contains all three — the body font is declared once in a part called docDefaults, each run of text carries its own size in half-points, and the page margins live on the section in twentieths of a point. A converter that asks you for them is one that never opened those parts.

What it does instead is throw your text into a blank page and lay it out again from scratch using whatever you picked. That's why the output never matches: it isn't a conversion, it's a re-typesetting. Your headings lose their style, your careful spacing collapses, and a two-column table becomes a run-on paragraph — not because a PDF can't hold those things, but because the tool never read that your document had them.

What a faithful conversion keeps

• Every run's font, point size, colour, bold, italic, underline and strikethrough — read from the document, not chosen for you.
• Paragraph alignment, left and right indents, first-line and hanging indents, space before and after, and line spacing.
• Headings with their styles, and real bulleted and numbered lists including their nesting.
• Tables with the column widths you actually set and the cell shading you applied.
• Inline images, embedded from their original bytes, and the page size and margins from the section.

A converter that reads the file can carry across almost everything you see on screen. The honest list of what should survive:

The one thing that genuinely can't cross over

Fonts are the honest limit. A PDF embeds its own fonts so it renders identically everywhere, and a browser tool works with the standard PDF font set rather than the hundreds installed on your machine. So Calibri and Aptos come out as a clean sans-serif, Cambria and Georgia as a serif, Consolas as a monospace. The layout holds to the pixel; the letterforms are a close cousin rather than an identical twin. For most documents nobody notices. For a brand deck where the exact typeface is the point, use Word's own Save as PDF, which embeds the real font.

The sharper limit is non-Latin text. The standard PDF fonts have no glyphs for Devanagari, Chinese, Arabic, Cyrillic or Greek — and when there's no glyph, the character doesn't warn you, it simply doesn't appear. A Hindi or Chinese document belongs in Word or LibreOffice, which embed the real font and keep the text. A good browser tool tells you when this is about to happen rather than shipping a page with holes in it.

Getting a clean result the first time

• Set your page size in Word before converting. A faithful tool uses the document's own size, so if you authored on Letter and want A4, change it at the source.
• Compress images first. A document full of phone photos makes a heavy PDF; use Word's Compress Pictures, or run the result through [Compress PDF](/pdf/compress/) afterwards.
• Fix wide tables at the source. A table wider than the text column is clipped, not shrunk — if it runs off the page in Word, it will in the PDF too.
• Check the fonts you rely on. If your document's whole identity is a specific typeface, or it's in a non-Latin script, Word's exporter is the right tool.

SnapToolz's [Word to PDF](/pdf/from-word/) reads the document's own formatting and leaves its restyle controls off by default — you convert first and only override if you actually want to change something. Going the other direction, [PDF to Word](/pdf/to-word/) rebuilds an editable .docx with the trade-offs that direction carries.