micromark extension to support MDX or MDX JS expressions

Usage no npm install needed!

<script type="module">
  import micromarkExtensionMdxExpression from 'https://cdn.skypack.dev/micromark-extension-mdx-expression';



Build Coverage Downloads Size Sponsors Backers Chat

micromark extension to support MDX (or MDX.js) expressions.

This package provides the low-level modules for integrating with the micromark tokenizer but has no handling of compiling to HTML: go to a syntax tree instead.

micromark-factory-mdx-expression and micromark-util-events-to-acorn are also maintained here.

When to use this

This package is already included in xdm and mdx-js/mdx (next).

You should probably use micromark-extension-mdx or micromark-extension-mdxjs instead, which combine this package with other MDX features. Alternatively, if you’re using micromark or mdast-util-from-markdown and you don’t want all of MDX, use this package.


This package is ESM only: Node 12+ is needed to use it and it must be imported instead of required.


npm install micromark-extension-mdx-expression


import * as acorn from 'acorn'
import {micromark} from 'micromark'
import {mdxExpression} from 'micromark-extension-mdx-expression'

// Agnostic (balanced braces):
const output = micromark('a {1 + 1} b', {extensions: [mdxExpression()]})


// Gnostic (JavaScript):
micromark('a {!} b', {extensions: [mdxExpression({acorn})]})


<p>a  b</p>
[1:5: Could not parse expression with acorn: Unexpected token] {
  reason: 'Could not parse expression with acorn: Unexpected token',
  line: 1,
  column: 5,
  source: 'micromark-extension-mdx-expression',
  ruleId: 'acorn',
  position: {
    start: { line: 1, column: 5, offset: 4 },
    end: { line: null, column: null }

…which is rather useless: go to a syntax tree with mdast-util-from-markdown and mdast-util-mdx-expression instead.


This package exports the following identifiers: mdxExpression. There is no default export.

The export map supports the endorsed development condition. Run node --conditions development module.js to get instrumented dev code. Without this condition, production code is loaded.


A function that can be called with options that returns an extension for micromark to parse expressions (can be passed in extensions).


Acorn parser to use (Acorn, optional).


Options to pass to acorn (Object, default: {ecmaVersion: 2020, sourceType: 'module'}). All fields can be set. Positional info (loc, range) is set on ES nodes regardless of acorn options.


Whether to add an estree field to mdxFlowExpression and mdxTextExpression tokens with the results from acorn (boolean, default: false). Note that expressions can be empty or be just comments, in which case estree will be undefined.


This extensions support both MDX and MDX.js. The first is agnostic to the programming language (it could contain Rust or so), the last is specific to JavaScript. To turn on gnostic mode, pass acorn.

There are two types of expressions: in text (inline, span) or in flow (block). They start with {.

Depending on whether acorn is passed, expressions are either parsed in several tries until whole JavaScript is found (as in, nested curly braces depend on JS expression nesting), or they are counted and must be balanced.

Expressions end with }.

For flow (block) expressions, optionally markdown spaces ( or \t) can occur after the closing brace, and finally a markdown line ending (\r, \n) or the end of the file must follow.

While markdown typically knows no errors, for MDX it is decided to instead throw on invalid syntax.

Here is an expression in a heading:

## Hello, {1 + 1}!

In agnostic mode, balanced braces can occur: {a + {b} + c}.

In gnostic mode, the value of the expression must be JavaScript, so
this would fail: {!}.
But, in gnostic mode, braces can be in comments, strings, or in other
places: {1 /* { */ + 2}.

The previous examples were text (inline, span) expressions, they can
also be flow (block):

  1 + 1

This is incorrect, because there are further characters:

  1 + 1

Blank lines cannot occur in text, because markdown has already split them in
separate constructs, so this is incorrect: {1 +


In flow, you can have blank lines:

  1 +



Unexpected end of file in expression, expected a corresponding closing brace for {

This error occurs if a { was seen without a } (source: micromark-extension-mdx-expression, rule id: unexpected-eof). For example:

a { b

Could not parse expression with acorn: Unexpected content after expression

This error occurs when there is more content after a JS expression (source: micromark-extension-mdx-expression, rule id: acorn). For example:

a {"b" "c"} d

Could not parse expression with acorn: $error

This error occurs if acorn crashes (source: micromark-extension-mdx-expression, rule id: acorn). For example:

a {var b = "c"} d


Two tokens are used, mdxFlowExpression and mdxTextExpression, to reflect flow and text expressions.

They include:

  • lineEnding for the markdown line endings \r, \n, and \r\n
  • mdxFlowExpressionMarker and mdxTextExpressionMarker for the braces
  • whitespace for markdown spaces and tabs in blank lines
  • mdxFlowExpressionChunk and mdxTextExpressionChunk for chunks of expression content



See contributing.md in micromark/.github for ways to get started. See support.md for ways to get help.

This project has a code of conduct. By interacting with this repository, organization, or community you agree to abide by its terms.


MIT © Titus Wormer