@mohtasham/md-to-docx

Convert Markdown to Microsoft Word (.docx) documents — in Node.js, in the browser, or straight from your terminal.

npm version npm downloads license types node

A TypeScript-first library and CLI that turns Markdown into production-ready Word documents: headings, tables, lists, images, code blocks with optional syntax highlighting, multi-section templates, per-section headers/footers, page numbering, TOC, and fine-grained style control.

Table of contents

• Highlights
• Installation
• Quick start
• CLI
• Programmatic usage
  • Browser
  • Node.js
  • React
• Features
  • Multi-section documents (template + sections)
  • Syntax-highlighted code blocks
  • Custom heading and paragraph alignment
  • Table of Contents styling
  • Text find-and-replace
  • RTL / bidirectional text
• Supported Markdown
• API reference
• Requirements
• Install as an agent skill
• Development
• Changelog
• Contributing
• License

Highlights

• Zero-config defaults — pass any Markdown string, get a valid .docx Blob back.
• First-class CLI — npx @mohtasham/md-to-docx input.md output.docx.
• TypeScript-native — fully typed options surface, including CodeHighlightTheme, Options, and DocumentSection.
• Multi-section documents — cover pages, per-section headers/footers, page numbering resets, mixed orientations, style overrides.
• Optional syntax highlighting — opt-in, powered by [lowlight](https://github.com/wooorm/lowlight); ships a GitHub-light theme and lets you override any token color.
• Works everywhere — Node.js (18+) and modern browsers; the package ships ESM with type declarations.
• Small public surface, stable API — only the root entrypoint is exported via package.json#exports.

Installation

npm install @mohtasham/md-to-docx
# or
pnpm add @mohtasham/md-to-docx
# or
yarn add @mohtasham/md-to-docx

Quick start

import { convertMarkdownToDocx } from "@mohtasham/md-to-docx";
import fs from "node:fs/promises";

const markdown = `
# Hello, Word

This document was generated from **Markdown** in TypeScript.

- Supports lists
- **Bold**, *italic*, ++underline++, ~~strikethrough~~
- Tables, blockquotes, images, and code blocks

\`\`\`ts
const greet = (name: string) => \`Hello, \${name}!\`;
\`\`\`
`;

const blob = await convertMarkdownToDocx(markdown);
await fs.writeFile("hello.docx", Buffer.from(await blob.arrayBuffer()));

CLI

Convert files without writing any code:

# Run without installing
npx @mohtasham/md-to-docx input.md output.docx

# Or install globally
npm install -g @mohtasham/md-to-docx
md-to-docx input.md output.docx

# Apply styling or multi-section config from a JSON file
md-to-docx input.md output.docx --options options.json
md-to-docx input.md output.docx -o options.json

# Help
md-to-docx --help

The --options JSON file accepts the same shape as the programmatic Options argument. Example:

{
  "documentType": "report",
  "style": {
    "fontFamily": "Trebuchet MS",
    "heading1Alignment": "CENTER",
    "paragraphAlignment": "JUSTIFIED"
  },
  "codeHighlighting": { "enabled": true }
}

For a multi-section document, use template + sections (the CLI ignores the positional markdown argument when sections is provided):

{
  "template": {
    "pageNumbering": { "display": "current", "alignment": "CENTER" }
  },
  "sections": [
    {
      "markdown": "# Cover\n\nPrepared for ACME",
      "footers": { "default": null },
      "pageNumbering": { "display": "none" }
    },
    {
      "markdown": "[TOC]\n\n# Body\n\nMain content…",
      "headers": { "default": { "text": "Main Section", "alignment": "RIGHT" } },
      "pageNumbering": { "start": 1, "formatType": "decimal" }
    }
  ]
}

Programmatic usage

Browser

import { convertMarkdownToDocx, downloadDocx } from "@mohtasham/md-to-docx";

const blob = await convertMarkdownToDocx("# Hello\n\nWorld.");
downloadDocx(blob, "hello.docx");

Node.js

import { convertMarkdownToDocx } from "@mohtasham/md-to-docx";
import fs from "node:fs/promises";

const blob = await convertMarkdownToDocx(await fs.readFile("input.md", "utf8"));
await fs.writeFile("output.docx", Buffer.from(await blob.arrayBuffer()));

React

import { useState } from "react";
import { convertMarkdownToDocx, downloadDocx } from "@mohtasham/md-to-docx";

export function MarkdownExporter() {
  const [markdown, setMarkdown] = useState("");

  const exportDocx = async () => {
    const blob = await convertMarkdownToDocx(markdown);
    downloadDocx(blob, "export.docx");
  };

  return (
    <>
      <textarea value={markdown} onChange={(e) => setMarkdown(e.target.value)} />
      <button onClick={exportDocx}>Export as DOCX</button>
    </>
  );
}

Features

Multi-section documents (template + sections)

Use template for shared section defaults and sections for explicit parts with their own markdown, headers, footers, page numbering, orientation, and style overrides.

const blob = await convertMarkdownToDocx("", {
  style: { fontFamily: "Trebuchet MS", paragraphSize: 24 },
  template: {
    page: {
      margin: { top: 1440, right: 1080, bottom: 1440, left: 1080 },
    },
    pageNumbering: { display: "current", alignment: "CENTER" },
  },
  sections: [
    {
      markdown: "# My Report\n\nPrepared for ACME",
      footers: { default: null },
      pageNumbering: { display: "none" },
      style: { paragraphAlignment: "CENTER", paragraphSize: 28 },
    },
    {
      markdown: "[TOC]\n\n# Executive Summary\n\n…",
      titlePage: true,
      type: "NEXT_PAGE",
      headers: {
        default: { text: "Executive Summary", alignment: "RIGHT" },
        first: { text: "Executive Summary (First)", alignment: "RIGHT" },
      },
      footers: {
        default: {
          text: "Page",
          pageNumberDisplay: "currentAndSectionTotal",
          alignment: "RIGHT",
        },
      },
      pageNumbering: { start: 1, formatType: "decimal" },
      style: { paragraphAlignment: "JUSTIFIED" },
    },
    {
      markdown: "# Appendix\n\n…",
      type: "ODD_PAGE",
      page: { size: { orientation: "LANDSCAPE" } },
      pageNumbering: { start: 1, formatType: "upperRoman" },
      style: { paragraphSize: 22 },
    },
  ],
});

Precedence (last wins): global style → template.style → per-section style. For headers / footers, each slot (default, first, even) can inherit, override, or be explicitly disabled with null.

Syntax-highlighted code blocks

Highlighting is opt-in; when disabled (the default) output is byte-identical to pre-highlighting versions. When enabled, each token becomes its own colored TextRun.

await convertMarkdownToDocx(markdown, {
  codeHighlighting: {
    enabled: true,
    showLanguageLabel: true,
    languages: ["typescript", "javascript", "python", "bash"],
    theme: {
      background: "0D1117",
      border: "30363D",
      default: "C9D1D9",
      languageLabel: "8B949E",
      keyword: "FF7B72",
      string: "A5D6FF",
      number: "79C0FF",
      comment: "8B949E",
      "title.function": "D2A8FF",
    },
  },
});

Theme keys map 1:1 to hljs-* token classes (without the hljs- prefix); values are RRGGBB hex strings without #. Reserved keys: default, background, border, languageLabel. Unknown or non-whitelisted languages fall back to the plain rendering path, so conversion never throws on an unsupported fence.

Custom heading and paragraph alignment

Each heading level and block type can be aligned independently:

await convertMarkdownToDocx(markdown, {
  style: {
    heading1Alignment: "CENTER",
    heading2Alignment: "RIGHT",
    heading3Alignment: "JUSTIFIED",
    paragraphAlignment: "JUSTIFIED",
    blockquoteAlignment: "CENTER",
  },
});

Set headingAlignment to provide a fallback for any level without its own override.

Table of Contents styling

Drop [TOC] on its own line in your markdown to render a clickable, auto-populated table of contents. Every TOC level is individually styleable:

await convertMarkdownToDocx(markdown, {
  style: {
    tocFontSize: 22,
    tocHeading1FontSize: 28, tocHeading1Bold: true,
    tocHeading2FontSize: 24, tocHeading2Bold: true,
    tocHeading3FontSize: 22,
    tocHeading4FontSize: 20, tocHeading4Italic: true,
    tocHeading5FontSize: 18, tocHeading5Italic: true,
  },
});

Text find-and-replace

Run string, regex, or functional replacements over the Markdown AST before conversion — they apply across every element type (headings, paragraphs, list items, table cells, etc.):

await convertMarkdownToDocx(markdown, {
  textReplacements: [
    { find: /oldText/g, replace: "newText" },
    { find: "Company Name", replace: "Acme Corp" },
    { find: /(\d+)/g, replace: (match) => `Number: ${match}` },
  ],
});

RTL / bidirectional text

await convertMarkdownToDocx(markdown, {
  style: {
    direction: "RTL",
    paragraphAlignment: "RIGHT",
  },
});

Supported Markdown

API reference

convertMarkdownToDocx(markdown, options?): Promise<Blob>

Converts Markdown text (or a multi-section template) to a DOCX Blob.

downloadDocx(blob, filename?): void

Browser-only helper that triggers a file download. Throws in non-browser environments. filename defaults to "document.docx".

Options

interface Options {
  documentType?: "document" | "report";
  style?: Style;
  template?: DocumentSection;
  sections?: DocumentSection[];
  codeHighlighting?: CodeHighlightOptions;
  textReplacements?: TextReplacement[];
}

Style

Text sizes

Spacing & layout

Alignment & direction

DocumentSection (for template and each entry in sections)

**pageNumbering:**

CodeHighlightOptions

TextReplacement

Errors

All conversion failures throw MarkdownConversionError (exported from the root). The error exposes a typed context field (e.g. { orientation, sectionIndex }) to make debugging easier.

Requirements

• Node.js ≥ 18 (ESM-only package)
• Browsers: any evergreen browser that supports ES2020 + Blobs

Install as an agent skill

For use in agent environments that speak the skills CLI:

# Quick add from GitHub URL
npx skills add https://github.com/mohtashammurshid/md-to-docx --skill md-to-docx

# From GitHub shorthand, pinned to an agent
npx skills add MohtashamMurshid/md-to-docx --skill md-to-docx --agent cursor --yes --full-depth

# From a local clone
npx skills add . --skill md-to-docx --agent cursor --yes --full-depth

# Discover before installing
npx skills add MohtashamMurshid/md-to-docx --list --full-depth

Development

git clone https://github.com/MohtashamMurshid/md-to-docx.git
cd md-to-docx
npm install

npm run build   # compile TypeScript to dist/
npm test        # run the Jest suite (4 suites, 45 tests)

Tests run offline against generated Word XML using JSZip. Set DEBUG_DOCX=1 to have tests also write .docx artifacts under test-output/ for manual inspection.

Changelog

See CHANGELOG.md for a detailed history.

Contributing

PRs and issues are welcome. A few guidelines:

1. Open an issue to discuss non-trivial changes before sending a patch.
2. Add or update tests under tests/ — the suite uses XML-level assertions on the generated DOCX, so changes to rendering should be verified at that level.
3. Keep the public API surface stable; internal refactors should not require bumping the major version.
4. Run npm run build && npm test before pushing.

License

MIT © Mohtasham Murshid Madani

Built on top of [docx](https://www.npmjs.com/package/docx), [unified](https://unifiedjs.com), [remark](https://github.com/remarkjs/remark), and [lowlight](https://github.com/wooorm/lowlight).