Upload a .docx and get Markdown: the headings, lists, links and tables come across, the fonts and margins do not.

Upload a .docx from Word, Google Docs or LibreOffice. What comes back is the document’s structure as Markdown — not its layout.

What comes across

A .docx is read for what each part of it is, not for how it looks on the page.

In the Word file	In the Markdown
Heading 1 to Heading 6	`#` to `######`
Bold, italic, strikethrough	`bold`, `italic`, `~~struck~~`
Bulleted and numbered lists	`-` and `1.` lists, two spaces in per level
Ticked and empty checkboxes in a list	`- [x]` and `- [ ]`
A table	A Markdown table, its first row as the header
A hyperlink	A link to the same address
A picture	Embedded in the document, with its alt text
Footnotes and endnotes	A numbered list at the end, `[1]` in the text
A line break (Shift+Enter)	A line ending in two spaces
A text box	Its words, after the paragraph it was anchored to

report.docx becomes report.md, and an asterisk or underscore in the text is escaped so it stays a character.

What is left behind

Styles that are not headings. Title, Subtitle, Quote, Heading 7 to 9 and your own styles become ordinary paragraphs.
Appearance. Fonts, sizes, colours, highlighting and underlining. Subscript and superscript lose their position, so H₂O arrives as H2O, and a monospaced font does not make code.
Comments, with no marker where they were.
Page furniture. Headers, footers, page numbers and page breaks.
Equations. Anything from Word's equation editor disappears and the sentence closes up.
Merged cells. A merged cell is written once and the cells after it shift left into the space it covered. Two paragraphs in one cell run together.
Working footnotes. The [1] and ↑ links point at anchors the Markdown lacks; these are not Markdown [^1] footnotes.
A working table of contents. Its entries arrive as links with page numbers, to bookmarks that are gone.

An example

A .docx with a Title-style line, a Heading 1, a paragraph with bold, italic, a footnote and a link, a nested list, a table and a photograph becomes this, the picture's base64 shortened:

Roof survey

# Findings

The north gutter is **blocked** and two tiles are *cracked*.[\[1\]](#footnote-1) Prices are on the [supplier’s page](https://example.com/tiles).

- Clear the gutter
  - North side only
- Replace two tiles

| Job | Cost |
| --- | --- |
| Gutter | £80 |
| Tiles | £240 |

![The blocked north gutter](data:image/png;base64,iVBORw0KGgo…)

1. Seen from the scaffold on 3 March. [↑](#footnote-ref-1)

"Roof survey" stays a plain paragraph because it was in the Title style.

Limits

.docx only, up to 50 MB. Several files dropped together count together and become one document, separated by rules.
The file is read in your browser, not uploaded. Nothing goes to an account unless you are signed in and press Save, and a saved document can be at most 4 MB.
Pictures travel inside the Markdown: about 1 MB each once encoded (a 750 KB file) and 2 MB in all for each .docx. One that does not fit becomes its alt text in italics, or nothing if it had none.
Lists nested under a numbered item are indented two spaces, which is not enough for Markdown: rendered, the sub-items leave their item. Nesting under bullets is fine.
A password-protected .docx cannot be read; remove the password in Word first.

Questions

How do I convert a Word document (.docx) to Markdown (.md)?

Drop the .docx here, check the preview and press Download .md. How to convert DOCX to Markdown compares this with Pandoc and with mammoth in your own code.

Can it convert an old .doc file?

No, only .docx. Save a .doc as .docx in Word or LibreOffice first; how to open a .docx file covers one that arrives with the wrong extension.

Can I convert a Google Doc?

Yes, by way of Word's format. In Google Docs choose File, Download, Microsoft Word, then drop that .docx here.

Why did my title not become a heading?

Word's Title style is not a heading style. Restyle the line as Heading 1 before converting, or put # in front of it afterwards.

What happens to comments and tracked changes?

Comments are dropped. Tracked changes come out as if all were accepted: inserted text stays and deleted text is gone, but text moved while tracking was on is lost from both places. Accept or reject every change in Word before you convert.

Which formatting is worth keeping?

Less than it looks. What not to keep from a .docx goes through what Markdown cannot say and which losses matter.

Word to Markdown