Word to Markdown

A Word to Markdown converter that runs in your browser

A .docx file is a zip of XML. This converter opens it with mammoth, a library built to turn Word documents into clean, semantic HTML — it reads paragraph styles rather than visual formatting, so a "Heading 2" becomes an h2 whatever font it uses. That HTML then goes through the same HTML-to-Markdown engine as our other tools (Turndown with GitHub-flavored Markdown support), plus a Word-specific cleanup pass. The whole DOCX to Markdown conversion happens on your device; nothing is uploaded.

The cleanup is where most online DOCX to Markdown converters fall short. Word's table of contents turns into a nested list of working links (with the page numbers removed), internal cross-references point at the right heading anchors, footnotes and endnotes become Markdown footnotes ([^1]), code styles and monospace paragraphs become fenced code blocks, and ☐/☒ checkboxes become task lists.

How to convert DOCX to Markdown

• Open the document. Drop a .docx onto the box or click to choose one. From Google Docs, use File → Download → Microsoft Word (.docx) first.
• Pick how images and tables come out. Embedded base64 keeps everything in one file; separate files keep the Markdown readable. HTML tables keep merged cells.
• Match your style. Choose # or underlined headings, the bullet character, and _ or * for emphasis. Options re-run the conversion instantly and are remembered.
• Copy or download. Check the rendered preview, tweak the text if you like, then copy it or save the .md (or a .zip with its images).

What carries over when you convert Word to Markdown

• Heading 1–6 and Title styles → # headings; bold, italic, strikethrough → **, _, ~~
• Bulleted and numbered lists, nested to any depth, with start numbers kept
• Tables → GitHub-flavored pipe tables (first row as header), or HTML tables with merged cells
• Hyperlinks, including links to headings inside the document
• Images → base64, separate files, or nothing; alt text is kept
• Footnotes and endnotes → [^n] references with their text at the end
• Quote and Block Text styles → > block quotes; code styles and monospace text → ` and ``` code
• Superscript and subscript → <sup> and <sub>

What doesn't: fonts, colors, sizes, alignment, headers and footers, comments, and tracked changes (only the current text is converted). Shapes, SmartArt, and equations may be skipped.

Pandoc: convert DOCX to Markdown from the command line

For batches or scripts, Pandoc is the standard DOCX to Markdown converter. Install it with brew install pandoc or winget install pandoc, then:

# github-flavored markdown, images saved to ./media
pandoc -f docx -t gfm --wrap=none --extract-media=./media input.docx -o output.md

# pandoc's own markdown flavor (grid tables, definition lists)
pandoc -f docx -t markdown --wrap=none input.docx -o output.md

# convert a whole folder of word documents
for f in *.docx; do pandoc -f docx -t gfm --wrap=none "$f" -o "${f%.docx}.md"; done

Tip: without --wrap=none, Pandoc hard-wraps paragraphs at 72 characters, which makes diffs of the Markdown noisy. With -t gfm, tables that can't be pipe tables (merged cells, several paragraphs in a cell) are written as HTML.

Python: mammoth, MarkItDown, and pypandoc

# pip install mammoth markdownify
import mammoth
from markdownify import markdownify

with open("input.docx", "rb") as f:
    html = mammoth.convert_to_html(f).value
open("output.md", "w").write(markdownify(html, heading_style="ATX"))

# pip install 'markitdown[docx]'   — microsoft's all-in-one converter
from markitdown import MarkItDown
print(MarkItDown().convert("input.docx").text_content)

# pip install pypandoc_binary
import pypandoc
pypandoc.convert_file("input.docx", "gfm", outputfile="output.md",
                      extra_args=["--wrap=none", "--extract-media=media"])

MarkItDown also has a CLI — markitdown input.docx -o output.md— and is popular for feeding Word files to LLMs. Mammoth's own Markdown output is deprecated, which is why the HTML → markdownify route is the usual recipe.

Tips for clean Markdown from Word

• Use Word's real heading styles (Home → Styles) rather than big bold text — they become proper # headings.
• Accept or reject tracked changes before converting; the converter only sees the current text.
• If a document is mostly a table of figures, try the HTML table mode and compare.
• For docs you'll paste into an LLM, drop images — base64 data is large and useless to a text model.
• Going the other way? The Markdown to Word converter writes styles that this tool reads back, so documents round-trip.

Frequently asked questions