Rich text with MDX

The rich-text type

Provides structured content which can embed custom templates that you define, much like the object type.

// .tina/schema.ts
import { defineSchema } from '@tinacms/cli'

export default defineSchema({
  collections: [
      fields: [
        // ...
          type: 'rich-text',
          label: 'Body',
          name: 'body',
          isBody: true,
          templates: [
              name: 'Cta',
              label: 'Call to Action',
              fields: [
                  type: 'string',
                  name: 'heading',
                  label: 'Heading',

An example

Given a markdown file like this:

## Hello, world!

This is some text

<Cta heading="Welcome" />

Results in the following response from the content API:

Notice the body response, it's a structured object instead of a string!


Since the value for rich-text is a structured object instead of a string, rendering it out to your page requires more work. We've provided a special component which is super lightweight and can used directly in your code.


// my-page.js
import { TinaMarkdown } from 'tinacms/dist/rich-text'

// The `props` here are identical to the respective template "fields"
const Cta = props => {
  return (

// Be sure to provide the appropriate components for each template you define
const components = {
  Cta: Cta

export default function MyPage = (props) => {
  return (
      <TinaMarkdown components={components} content={props.body} />
type TinaMarkdown = ({
  // The rich-text data returned from the content API
  content: TinaMarkdownContent
   * Any templates provided in the rich-text field.
   * Optionally, most elements (ex. <a>) can also
   * be overridden
  components?: Components<{}>
}) => JSX.Element

Differences from other MDX implementations

If you've worked with MDX before, you know that there's usually a compilation step which turns your .mdx file into JavaScript code. This works really well for developers who can access their files directly, but it creates problems when editing content from a rich-text interface. With Tina, we're leveraging a subset of MDX to enable what's most important to content editors, and in doing so it's necessary to narrow the scope of what's supported:

All JSX must be registered as a template

In the above example, if you failed to add the Cta template in your schema definition, you would receive an error:

Found unregistered JSX or HTML: <Cta>. Please ensure all structured elements have been registered with your schema.

All content must be serializable

When we say serializable, we mean that they must not be JavaScript expressions that would need to be executed at any point.

  • No support for import/export
  • No support for JavaScript expressions (eg. const a = 2, console.log("Hello"))

For example:

## Today is {new Date().toLocaleString()}

This expression will be ignored, instead register a "Date" template:

## Today is <Date />

Then you can create a Date component which returns new Date().toLocaleString() under the hood.

You must supply the appropriate components yourself

Traditionally, MDX will compile whatever components you import so they're in scope when it's time to render your content. But with Tina, MDX is completely decoupled from your JavaScript so it's up to you to ensure that for every template in your rich-text definition, there's an equivalent component in your <TinaMarkdown> component.