Guide · DOCX → HTML
What survives when Word becomes HTML
"Keep the formatting" means two different things, and which one you mean decides whether you'll be happy with the result.
Structure — headings, lists, tables, emphasis — comes through. Appearance — fonts, colours, margins, page breaks — does not, and that's a deliberate choice rather than a missing feature.
Structure survives, appearance doesn't
A .docx stores two separate things about every paragraph: what it is, and what it looks like. "Heading 2" is what it is. 16pt Calibri bold with 6pt space after is what it looks like.
The converter reads the first and throws away the second. A Heading 2 becomes an <h2> — not a <p> with a font size on it. On your site it then picks up your own heading style, which is almost certainly what you wanted.
The alternative is what Word's own Save As Web Page does: a few thousand lines of mso- styles that reproduce the page exactly and can't be edited or restyled afterwards.
Style your headings, don't just embolden them
This is the one thing worth doing in Word before you convert, and it's the difference between good output and flat output.
Text that merely looks like a heading — bigger, bold, manually sized — is stored as a normal paragraph. It converts to <p>, because that's genuinely what it is. Nothing can recover the intent from the font size.
- 01In Word, put the cursor in a heading and check the Styles gallery. If it says Normal, that's the problem.
- 02Apply Heading 1, 2 or 3 instead. It'll change how it looks; restyle the heading style itself if you don't like it.
- 03Do the same for lists: use the list buttons rather than typing "1." and a tab. Manually typed numbers convert to literal text, not to an <ol>.
- 04Convert, then check the output for <h2> and <ul>. If you see a wall of <p>, the document was never structured to begin with.
What comes through
Heading levels as <h1> through <h6>, taken from Word's styles. Bold, italic and strikethrough as <strong>, <em> and <s>. Superscript and subscript. Links, with Google's and Office's tracking parameters stripped off.
Lists as real <ul> and <ol>, nested up to five levels. Tables as <table>, including merged cells: horizontal merges come through as colspan and vertical ones as rowspan.
Blockquotes from the Quote style. Paragraphs styled Code or Preformatted as <pre><code>. Captions as <p class="caption">.
What gets dropped, and why
Fonts, sizes, colours, highlighting, alignment, indentation, line spacing, page breaks, headers and footers, margins. All of it is presentation that belongs to a page, and HTML isn't a page.
Two drops are worth knowing about because they're easy to miss. Underlined text loses its underline — on the web an underline means a link, so Word's underline maps to nothing rather than to something misleading. And blank spacer paragraphs disappear, because they're layout, not content.
Track changes and comments are dropped: you get the final text, not the editing history. Text boxes, SmartArt and charts don't survive — only their text, if they contain any.
Images: three choices
Old .doc files are the exception: images can't be recovered from that format at all, no matter which option you pick. Everything else — text, headings, tables, bold — comes through.
- 01Inline as base64 gives you one self-contained .html file you can email or double-click. It's about a third bigger than the images were.
- 02Separate files gives you a zip: the HTML plus an images/ folder, with every <img> already pointing at the right path. Use this for anything going on a real site.
- 03Dropped leaves the <img> tags in place with an empty src, so you'll see broken-image placeholders. Delete them, or use one of the other two.
Set your heading styles first, then convert. Files are read in your browser and never uploaded, so you can try it on something confidential.
DOCX → HTML