Command line

word2html [-h] [--version] [--strict] [--timeout SECONDS]
          [--max-memory SIZE] [--max-file-size SIZE] path

Options

Option	Default	What it does
`path`	Required	The `.docx` file to convert. The extension isn't case-sensitive.
`--strict`	Off	Refuse documents with unsafe links or hidden text instead of warning. See Security.
`--timeout SECONDS`	`120`	Stop pandoc after this long. Decimals are allowed.
`--max-memory SIZE`	`2G`	Memory limit for pandoc. The minimum is `16M`.
`--max-file-size SIZE`	`200M`	The largest input file to accept.
`--version`		Show the version and exit.
`-h`, `--help`		Show help and exit.

SIZE is a whole number with an optional unit: K, M, or G, such as 512M or 2G. A number alone means bytes. Units are powers of 1024, and 512MB and 512MiB both mean the same as 512M.

Exit codes

Code	Meaning	Output file
`0`	Converted. Any warnings were printed to stderr.	Written
`1`	Invalid input, a limit was exceeded, or pandoc or the file write failed.	Not changed
`2`	Bad command line.	Not changed
`3`	Refused for safety.	Not changed

The output file

Location: next to the input, with the same name and an .html extension. docs/Report.v2.DOCX becomes docs/Report.v2.html.
Overwriting: an existing file with that name is replaced without asking, keeping its permissions.
Atomic: the new file is written separately and then swapped into place. If conversion fails or is refused, an existing file stays exactly as it was.
Symlinks: if the output path is a symlink, the link itself is replaced and its target isn't touched.
Page title: the text of a paragraph in Word's Title style if there is one, and otherwise the filename.

The success message goes to stdout. Warnings and errors go to stderr, so you can separate them in scripts.

What's kept and what isn't

In the Word document	In the HTML
Text, headings, lists, tables, links, footnotes, highlights	Kept
Embedded images, PDFs, video, audio	Embedded in the HTML, so output can be larger than the `.docx`
Linked images, stored outside the document	Dropped with a warning, and never fetched
Tracked insertions	Published as if accepted
Tracked deletions	Dropped
Comments	Dropped
Paragraphs in Word's Title, Author, and Date styles	Kept, and also used for the page's metadata
Document properties, set in File › Info	Not published
Hidden text	Published as visible text, with a warning. See Security.

Styling comes from pandoc's default CSS rather than the Word document, so fonts, colours, and page layout won't match the original.

Recipes

Convert a folder of documents

Use --strict and track failures, so one bad document doesn't go unnoticed:

status=0
for doc in ./*.docx; do
  word2html --strict "$doc" || status=1
done
exit "$status"

The ./ matters: it stops a file named something like --strict.docx from being read as an option.

Convert a file whose name starts with a dash

word2html ./-draft.docx
word2html -- -draft.docx

Act on the result in a script

word2html --strict "$doc"
case $? in
  0) echo "converted" ;;
  3) echo "refused: needs a human to review it" ;;
  *) echo "failed" ;;
esac

Convert untrusted documents on a shared machine

word2html --strict --timeout 30 --max-memory 512M --max-file-size 25M upload.docx

Then read Converting documents you don't trust, because these checks aren't a complete sanitizer.

Convert a very long document

word2html --max-memory 8G --timeout 600 book.docx