Convert HTML to a Word .docx file
Paste HTML or drop an .html file and get a clean Word document with real styles, working lists and converted tables — processed locally.
- Runs in your browser
- Files are never uploaded
- Free, no account
The old trick, and why it is bad
The advice you will find everywhere is to rename an .html file to
.doc. Word will open it, because it still supports an ancient HTML-based
format. What you get is not a Word document: it is HTML in a Word-shaped wrapper. Styles do
not exist, the navigation pane is empty, "Save As .docx" produces a mess, and other
applications refuse it outright.
This page produces a real OOXML .docx instead — the same zip-of-XML structure Word writes when you press Save.
How the conversion runs
- HTML is parsed and reduced to structure. The markup goes through a Markdown conversion step, which is a deliberate way of discarding everything presentational and keeping what carries meaning.
- Tables are handled explicitly. A dedicated rule walks every
<table>, reads its rows and cells, and emits a pipe table so the structure is not lost in the reduction. - The result is written as OOXML. Headings, lists, quotes, code and tables each map to Word's own built-in style ids, so the document inherits whatever theme the reader has applied.
The intermediate step is why the output is clean. A direct HTML-to-Word mapping tries to
preserve <span style="..."> and ends up with hundreds of one-off
character formats; going via structure throws all of that away first.
What is kept and what is dropped
| HTML | Outcome |
|---|---|
h1–h6 | Word Heading 1–6, with outline levels |
strong, b, em, i, del | Character formatting |
ul, ol, nested lists | Real Word lists that renumber when edited |
table | A Table Grid table with a bold header row |
a href | A clickable hyperlink with a relationship entry |
pre, code | Preformatted paragraphs and inline code style |
blockquote | Word's Quote style |
| CSS, classes, inline styles | Discarded on purpose |
img | A text placeholder — nothing is fetched |
script, iframe, svg | Removed |
Where the HTML usually comes from
- An email or newsletter export. Expect the table-based layout scaffolding to disappear, which is almost always what you want.
- A CMS or documentation site export. These convert well — the markup tends to be semantic already.
- A saved web page. The whole page comes through, navigation and footer included. Deleting those two blocks in Word afterwards is faster than any heuristic that tries to guess where the article starts.
- A rich-text editor's output. Usually very clean, since editors emit simple semantic markup.
- Rendered Markdown. Skip a step and use Markdown to Word on the source instead.
Images, and why they are not embedded
An <img src="https://…"> is a reference to a file on someone else's
server. Embedding it would require this page to make a network request, and the whole site
is built so that it cannot — the content security policy permits no third-party connections
at all.
Rather than fail quietly, each image becomes an italic [image: alt text] marker in
place, so you can see where they were and drop the real ones in. If your HTML uses
data: URIs, the same applies — supporting those without also supporting remote
ones would be an inconsistent rule that is easy to get wrong.
Nothing is uploaded, and that matters here
HTML pasted into a converter is frequently an internal wiki page, a client-facing report pulled out of a CRM, or an email thread. The parsing, the reduction and the zipping all happen in this browser tab; the .docx is assembled in memory and saved straight to your downloads folder.
Related: HTML to Markdown if you want the intermediate form, and Text to Word for plain text input.
Questions
How is the HTML interpreted?
It is converted to Markdown first, using the same rules as our HTML to Markdown tool, and then written out as OOXML. That intermediate step strips presentational markup and leaves the structure, which is what Word actually wants.
Do CSS styles carry over?
No. Inline colours, fonts and layout are dropped on purpose. What survives is meaning — headings, emphasis, lists, tables, links and code — which is then rendered with Word’s own styles so it matches your template.
What about images?
Images are noted in place as a bracketed caption rather than embedded, because an <img> usually points at a remote URL and this page makes no network requests of any kind. Paste the image into Word yourself afterwards.
Can I convert a whole saved web page?
Yes — drop the .html file in. Expect a lot of navigation and footer text to come along with it; deleting those in Word afterwards is quicker than trying to guess which parts of a page are the article.