Skip to main content
Deno 2 is finally here 🎉️
Learn more

DOCX Markup Language

This is a Deno module for making .docx files from scratch or from an existing DOCX/DOTX template. It will probably be published for NodeJS soon™ too.

You could use docxml to:

  • Create an MS Word file without ever opening MS Word
  • Create a parameterized template file, and render it to DOCX with your data parameters
  • Convert JSON, XML or other data structures to DOCX
  • Parse content from an existing DOCX file
  • Extract style definitions from a DOTX/DOCX file

docxml is designed to be used in vanilla JavaScript using class component instances, or using JSX if you’re on Deno or want to use NodeJS and a transpiler like Babel:

const para = new Paragraph(
    { alignment: 'center' },
    new Text({}, 'I want a cookie')
);
/** @jsx Docx.jsx */
const para = (
    <Paragraph alignment="center">
        <Text>I want a cookie</Text>
    </Paragraph>
);

👉 More on using class components

👉 More on using JSX

docxml is also designed to be used from scratch/entirely programmatically, or using a more ergonomic API to transform from an XML document. Both modes work equally well with vanilla JS or JSX.

await Docx
    .fromJsx(
        <Paragraph alignment="center">
            <Text>I want a cookie</Text>
        </Paragraph>,
    )
    .toFile('example-1.docx');
await Docx.fromNothing()
    .withXmlRule(
        'self::text()',
        ({ node }) => <Text>{node.nodeValue}</Text>
    )
    .withXmlRule(
        'self::p',
        ({ traverse, node }) => (
            <Paragraph alignment={node.getAttribute('align')}>
                {traverse()}
            </Paragraph>
        )
    )
    .withXml(`<p align="center">I want a cookie</p>`, {})
    .toFile('example-2.docx');

👉 More on XML rendering rules

👉 Go to the API docs that Deno generates for docxml

Features

To great or small extend, the following features work in the current version of docxml. Some items are not ticked off yet – they are not available, but hopefully soon.

👉 See code examples of some or the more intricate features

API features:

  • 100% typed
  • Asynchronous components
  • Component composition

Custom styles:

  • Font size and color
  • Bold, italic, underline styles, strike-through
  • Subscript, superscript, small caps
  • Paragraph spacing and indentation
  • Left/right/center/justified alignment
  • Numbering
  • Aligning text on tabs
  • Font family
  • Embed TTF in the DOCX file

References:

  • Cross references
  • Table of contents

Tables:

Images:

  • From any UInt8Array source
  • Alternative and title text
  • Width and height

Sections:

  • Width and height
  • Orientation
  • Page headers & footers

Comments:

  • Point comment
  • Range comment
  • Comment reply

Change tracking:

  • Text additions and deletions
  • Style changes
  • Table row additions and deletions

Differences with actual MS Word DOCX

Obviously docxml is a TypeScript project, which is already very different from how you would normally interact with a DOCX document. More meaningfully however, docxml is meant to make writing DOCX easier than going straight to OOXML. For example;

  • All sizes are of type Length, which means it doesn’t matter wether you input them as points, centimeters, inches, 1/2, 1/8th or 1/20th points, English Metric Units, and so on.
  • The JSX pragma will try to correct components that would lead to invalid XML structures, by splitting the parents of invalidly placed components recursively until the new position is valid. Moreover, string content in unexpected places is automatically wrapped in <Text> when using JSX. This makes the configuration of a new DOCX a little more forgiving.
  • Using the <Image> or <Comment> components will automatically create all required relationships etc.
  • Some of the words have changed, generally speaking docxml is more verbose than the DOCX verbiage.
  • Generally speaking docxml prefers formal (JS) references over references-by-identifier. In those cases the identifiers are randomly generated for you when the .docx file is written.
  • Especially in tables and images, a lot of formatting details are automatically applied. In a lot of cases there is no API yet to change them.

For contributors

This project uses unit tests and linting for quality control. To lint, both Deno’s own linting as well as ESLint are used. Please run both of the following commands to ensure that a GitHub Action does not fail later.

# Once
npm install

# Run all unit tests
deno task test


# Run all linting
deno task lint