How do you convert HTML to Word, and what actually survives?

Updated: 2026-10-14 · Dawwerni team

To convert HTML to Word, open the .html file in Word and save it as .docx, run pandoc page.html -o page.docx, import it into Google Docs, or use a free in-browser converter. All of them keep the structure: headings, paragraphs, lists, tables, links and images. None of them reproduce CSS layout such as columns, cards or backgrounds.

Structure converts, layout does not

A Word document and a web page are built on different models. Word thinks in a flow of paragraphs, each with a style (Normal, Heading 1, List Paragraph), plus tables and pictures anchored in that flow. A web page is a tree of boxes positioned by CSS: flexbox, grid, absolute positioning, background images, shadows, gradients. Converting HTML to Word means translating the tree into the flow, and only the parts that have a Word equivalent make the trip.

That is why the honest answer to "can I convert this HTML to Word?" is: yes, the content and its structure, but not the look. If you need the exact look, a PDF is the right format (see our guide to converting HTML to PDF for free). If you need to edit the text, comment on it, or paste it into a company template, Word is the right format and the free HTML to Word tool or one of the methods below will get you there.

Method 1: open the HTML file directly in Word

Microsoft Word can open .html files. In Word, choose File, then Open, browse to the file (pick "All files" if it is hidden by the filter) and open it. Word renders the page with its own HTML engine, then you use File, Save As and pick Word Document (.docx).

It is the fastest route if you already have Word, but expect a rough result on modern pages. Word's HTML engine is old: it does not understand flexbox or grid, ignores most modern CSS, does not run JavaScript, and often keeps odd leftovers such as fixed widths or web fonts that are not installed. Text and tables come through; everything that depends on CSS layout or scripts does not.

1. Save the page as a single .html file (images in the same folder if they use relative paths).
2. In Word, File > Open and select the .html file.
3. Check headings: if they look like headings but are only bold text, apply Heading 1/2/3 styles.
4. File > Save As > Word Document (.docx). Saving as .docx matters; otherwise Word keeps it as a web page.

Method 2: Pandoc, the cleanest result for developers

Pandoc is a free, open-source command-line converter. It reads the HTML structure, ignores CSS entirely, and writes a tidy .docx that uses real Word styles. That makes it the best choice when you want a clean document to restyle, not a copy of the web page.

# basic conversion
pandoc page.html -o page.docx

# use your own Word styles (fonts, colours, heading sizes)
pandoc -o custom-reference.docx --print-default-data-file reference.docx
# open custom-reference.docx in Word, edit the styles, save, then:
pandoc page.html --reference-doc=custom-reference.docx -o page.docx

What Pandoc does not do

Pandoc does not run JavaScript, so a page whose content is rendered by scripts (a React app, a chart drawn by Chart.js) arrives empty or incomplete. A canvas chart has no content to convert. Pages produced by AI tools often load Tailwind and Chart.js from a CDN; Pandoc will keep the text and tables but the charts disappear. If the document needs those charts, either export them as images first or use a converter that renders the page in a browser.

Method 3: Google Docs import

If you do not have Word, Google Drive can do the job. Upload the .html file to Drive, right-click it, choose Open with, then Google Docs. Drive converts it into a Google Doc. From there, File, Download, Microsoft Word (.docx) gives you the Word file.

The result is similar to Word's own import: headings, lists, links and tables are fine, layout is flattened, scripts do not run. Images referenced with relative paths will be missing because Drive only sees the single file you uploaded; images hosted on public URLs usually come through.

Method 4: an in-browser converter that renders the page first

The methods above read the HTML source. That works for hand-written documents but fails for pages that build themselves in the browser, which is exactly what most AI-generated HTML does. Our free HTML to Word converter takes the opposite approach: it runs the page in a sandboxed iframe in your browser, waits for fonts, images and an adjustable delay so CDN styles and charts finish drawing, then walks the rendered result.

It turns headings, paragraphs, bold, italic, links, lists, tables and code blocks into real editable Word text, and inserts images and charts as pictures, so a Chart.js bar chart is kept as an image instead of vanishing. Everything happens on your machine: the file is not uploaded, there is no sign-up and no watermark. You can drop one or more .html files, paste code, or drop a .zip that contains the HTML with its images and CSS.

It has the same fundamental limit as every other method: it converts structure, not visual layout. Two-column grids become stacked paragraphs, cards lose their borders and shadows, coloured backgrounds disappear. Images on other websites that block cross-origin access can also be missing in the captured pictures.

Which method should you use?

For static documents and batch jobs, Pandoc is hard to beat. For a page made by an AI assistant with charts and CDN styles, a converter that renders it first is easier; the AI HTML converter offers Word next to PDF, PowerPoint, images and Markdown if you are not sure yet which format you need.

Write Word-friendly HTML (or ask the AI to)

The best conversion starts with good HTML. If you control the source, or you are asking Claude, ChatGPT or Gemini to generate it, a few habits make the .docx far cleaner.

• Use real headings in order (h1, then h2, then h3), not divs styled to look big.
• Use real lists (ul, ol, li) instead of paragraphs that start with a dash or emoji.
• Put tabular data in a table with a thead row of th cells; avoid grids of divs.
• Give every image an alt text and a reasonable size; very wide images get scaled to the page width.
• Keep important text as text, not inside images or canvas.
• Avoid content that only appears on hover, in tabs or in accordions; closed sections may be skipped.
• Use pre and code for code, so it stays monospaced.

Prompt snippet to add when asking an AI for the HTML:
"Use semantic HTML: one h1, h2/h3 for sections, real ul/ol lists,
data in <table> with <thead>, no content hidden in tabs or accordions.
Keep it readable without CSS, because I will convert it to Word."

After conversion: a 5-minute cleanup in Word

1. Open the navigation pane (View > Navigation Pane) and check that the heading tree matches the page.
2. Select all and apply your company font, or use Design > Themes to restyle in one go.
3. Check tables: Table Design > pick a style, then Layout > AutoFit > AutoFit Window.
4. Resize any picture that overflows the margins.
5. Insert a table of contents (References > Table of Contents) if the document is long; it works because the headings are real styles.

If you also need a version that looks exactly like the original page, export a PDF alongside the Word file. When the PDF comes out wrong, this list of HTML to PDF fixes covers fonts, dark backgrounds and page breaks, and if the page is really a slide deck, converting HTML slides to PPTX is the better path. For plain-text workflows such as a wiki or a README, HTML to Markdown is lighter than Word.