🚀 Effortlessly convert Markdown Abstract Syntax Trees (MDAST) to DOCX.
✅ MDAST to DOCX conversion — Supports standard Markdown elements
✅ Footnotes handling — Converts Markdown footnotes to DOCX format
✅ Tables support — Converts Markdown tables into DOCX tables
✅ Hyperlink support — External and internal hyperlinks are now fully functional
✅ New Plugin Architecture — Extend and customize DOCX output with plugins
✅ Customizable image handling — Images are now supported via @m2d/image plugin
Note: Images are no longer supported by default. To enable image support, use the image plugin from
@m2d/imageormdast2docx/pluginsexplicitly.
This change helps us keep the library lean and extensible by community via plugins.
Converts markdown content/artifacts generated by Generative AI to docx for further processing/record-keeping/sharing.
Modern Generative AI tools often produce structured content in Markdown (MD) format. However, converting this AI-generated Markdown into professional DOCX documents requires an additional transformation step. This is where MDAST to DOCX comes in.
- AI tools generate content in Markdown.
- Unified.js & Remark Ecosystem convert Markdown into Markdown Abstract Syntax Tree (MDAST).
- MDAST to DOCX processes this structure and exports a well-formatted DOCX document.
- AI-Generated Reports – Convert AI-generated Markdown reports into polished DOCX files.
- AI-Powered Content Export – Transform AI-generated blogs, research papers, and structured content into a DOCX format for easy sharing and editing.
- Client & Server Flexibility – Works both in browser-based AI tools and high-performance server environments.
- Works on both Client & Server – Choose to offload processing to the client for better scalability or run it on the server for batch conversions.
- Optimized & Extensible – Use only the plugins you need, keeping your AI workflows lightweight.
By integrating MDAST to DOCX, AI applications can seamlessly export Markdown-based content into DOCX, making it more accessible, editable, and shareable. 🚀
pnpm add mdast2docxpnpm add @m2d/core @m2d/html @m2d/image @m2d/math @m2d/table @m2d/list @m2d/mdastimport { toDocx } from "mdast2docx";
import { unified } from "unified";
import remarkParse from "remark-parse";
import remarkMmd from "remark-mmd"; // Assumed MMD plugin
const markdown = `
# Sample Document
This is a **bold** text and _italic_ text.
> A blockquote example
- List Item 1
- List Item 2
[Click Here](https://example.com)
This is a footnote reference[^1].
[^1]: This is the footnote content.
`;
const mdast = unified().use(remarkParse).use(remarkMmd).parse(markdown);
(async () => {
const docxBlob = await toDocx(mdast, { title: "My Document" }, {});
const url = URL.createObjectURL(docxBlob as Blob);
const link = document.createElement("a");
link.href = url;
link.download = "document.docx";
link.click();
URL.revokeObjectURL(url);
})();| Parameter | Type | Description |
|---|---|---|
astInputs |
Root | { ast: Root; props?: ISectionProps }[] |
Single or multiple MDAST trees with optional section properties. |
docxProps |
IDocxProps |
General document properties (title, styles, metadata). |
defaultSectionProps |
ISectionProps |
Default properties for document sections. |
outputType (optional) |
"blob" (default) | "buffer" | "base64" |
Format of the generated DOCX document. |
📌 Returns: A Promise resolving to a DOCX document in the chosen format.
| Markdown Syntax | Supported in DOCX |
|---|---|
Headings # H1 to ###### H6 |
✅ |
| Paragraphs | ✅ |
Bold **text** & Italics _text_ |
✅ |
Blockquotes > quote |
✅ |
| Lists (ordered & unordered) | ✅ (ordered lists via listPlugin) |
Links [text](url) |
✅ |
Images  |
✅ (via imagePlugin) |
Code Blocks `code` |
✅ |
Footnotes [^1] |
✅ |
| Tables | ✅ (via tablePlugin) |
| Hyperlinks (external & internal) | ✅ |
| Latex Math | ✅ (via mathPlugin) |
| HTML Tags | ✅ (via htmlPlugin) |
You can customize the DOCX output using ISectionProps and IDocxProps.
const docxProps = {
title: "Styled Document",
author: "John Doe",
};
const sectionProps = {
page: { margin: { top: 1000, bottom: 1000, left: 1200, right: 1200 } },
};
const docxBlob = await toDocx(mdast, docxProps, sectionProps);Plugins allow extending functionality like adding image or table support.
import { toDocx } from "@m2d/core";
import { imagePlugin } from "@m2d/image";
import { tablePlugin } from "@m2d/table";
import { listPlugin } from "@m2d/list";
import { mathPlugin } from "@m2d/math";
const downloadDocx = () => {
toDocx(
mdast,
{},
{ plugins: [imagePlugin(), tablePlugin(), listPlugin(), mathPlugin()] },
"blob",
).then(blob => {
const url = URL.createObjectURL(blob as Blob);
const link = document.createElement("a");
link.href = url;
link.download = "my-document.docx";
link.click();
URL.revokeObjectURL(url);
});
};📖 More details:
AI tools often generate Markdown as output (e.g., reports, documentation, summaries). This library makes it easy to convert those Markdown-based outputs into rich DOCX documents.
- Supports both client-side and server-side usage
- Enables offloading document generation to users' browsers or running high-performance server renderers
- Combine it with
unifiedandremarkto parse Markdown into MDAST, and then convert to DOCX seamlessly
We welcome contributions! To get started:
- Fork the repository
- Create a feature branch (
git checkout -b feature-new) - Commit your changes (
git commit -m "Add new feature") - Push to GitHub (
git push origin feature-new) - Open a Pull Request 🚀
The community can create plugins to extend the functionality of this library. For inspiration, check out the existing plugins in the lib/src/plugins folder.
This project is licensed under MPL-2.0. See the LICENSE file for details.
Thanks to the docx.js and unified.js ecosystems for making this possible.
⭐ Star this repository if you find it useful!
❤️ Support our work by sponsoring.
Made with 💖 by Mayank Kumar Chaudhari