vitepress Documentation

Vite & Vue powered static site generator.

18,155 stars TypeScript TypeScript 1 file · ~42,827 tokens #docs-generator#hacktoberfest#markdown#ssg#vite#vue
RAW Doc

1. Project Overview & Quickstart (vuejs/vitepress)

File: README.md

VitePress 📝💨





VitePress is a Vue-powered static site generator and a spiritual successor to VuePress, built on top of Vite.

Documentation

To check out docs, visit vitepress.dev.

Changelog

Detailed changes for each release are documented in the CHANGELOG.

Contribution

Please make sure to read the Contributing Guide before making a pull request.

License

MIT

Copyright (c) 2019-present, Yuxi (Evan) You

Special Thanks

This project would not be possible without the support of these amazing companies and tools.

VoidZero The company behind Vite, Vitest, Rolldown, and Oxc. VoidZero has been a long-time sponsor of VitePress. VitePress itself is built on top of this ecosystem of high-performance JavaScript tooling.
StackBlitz Creators of WebContainers and a browser-based development platform. Their technology powers the VitePress playground and provides instant preview releases for our pull requests and commits.
Algolia An AI-powered search platform that provides fast, scalable search. Algolia powers DocSearch for VitePress. Their team has been a long-time partner in building and maintaining our search integration.
BrowserStack A cloud-based testing platform for websites and mobile apps. BrowserStack provides the cross-browser and real-device infrastructure used to ensure VitePress works consistently across all browsers.
Netlify A platform for building, deploying, and hosting modern web projects. Netlify powers our website deployments with their generous open source plan and provides deploy previews for the pull requests.

File: docs/en/guide/asset-handling.md



Asset Handling

Referencing Static Assets

All Markdown files are compiled into Vue components and processed by Vite. You can, and should, reference any assets using relative URLs:

md

You can reference static assets in your markdown files, your *.vue components in the theme, styles and plain .css files either using absolute public paths (based on project root) or relative paths (based on your file system). The latter is similar to the behavior you are used to if you have used Vite, Vue CLI, or webpack's file-loader.

Common image, media, and font filetypes are detected and included as assets automatically.

::: tip Linked files are not treated as assets
PDFs or other documents referenced by links within markdown files are not automatically treated as assets. To make linked files accessible, you must manually place them within the public directory of your project.
:::

All referenced assets, including those using absolute paths, will be copied to the output directory with a hashed file name in the production build. Never-referenced assets will not be copied. Image assets smaller than 4kb will be base64 inlined - this can be configured via the vite config option.

All static path references, including absolute paths, should be based on your working directory structure.

The Public Directory

Sometimes you may need to provide static assets that are not directly referenced in any of your Markdown or theme components, or you may want to serve certain files with the original filename. Examples of such files include robots.txt, favicons, and PWA icons.

You can place these files in the public directory under the source directory. For example, if your project root is ./docs and using default source directory location, then your public directory will be ./docs/public.

Assets placed in public will be copied to the root of the output directory as-is.

Note that you should reference files placed in public using root absolute path - for example, public/icon.png should always be referenced in source code as /icon.png.

Base URL

If your site is deployed to a non-root URL, you will need to set the base option in .vitepress/config.js. For example, if you plan to deploy your site to https://foo.github.io/bar/, then base should be set to '/bar/' (it should always start and end with a slash).

All your static asset paths are automatically processed to adjust for different base config values. For example, if you have an absolute reference to an asset under public in your markdown:

md

You do not need to update it when you change the base config value in this case.

However, if you are authoring a theme component that links to assets dynamically, e.g. an image whose src is based on a theme config value:

vue

In this case it is recommended to wrap the path with the withBase helper provided by VitePress:

vue
<script setup>
import { withBase, useData } from 'vitepress'

const { theme } = useData()
</script>

<template>
  
</template>

File: docs/en/guide/cms.md


outline: deep

Connecting to a CMS

General Workflow

Connecting VitePress to a CMS will largely revolve around Dynamic Routes. Make sure to understand how it works before proceeding.

Since each CMS will work differently, here we can only provide a generic workflow that you will need to adapt to your specific scenario.

  1. If your CMS requires authentication, create an .env file to store your API tokens and load it so:

    js
    // posts/[id].paths.js
    import { loadEnv } from 'vitepress'
    
    const env = loadEnv('', process.cwd())
  2. Fetch the necessary data from the CMS and format it into proper paths data:

    js
    export default {
      async paths() {
        // use respective CMS client library if needed
        const data = await (await fetch('https://my-cms-api', {
          headers: {
            // token if necessary
          }
        })).json()
    
        return data.map(entry => {
          return {
            params: { id: entry.id, /* title, authors, date etc. */ },
            content: entry.content
          }
        })
      }
    }
  3. Render the content in the page:

    md
    # {{ $params.title }}
    
    - by {{ $params.author }} on {{ $params.date }}

Integration Guides

If you have written a guide on integrating VitePress with a specific CMS, please use the "Edit this page" link below to submit it here!


File: docs/en/guide/custom-theme.md



Using a Custom Theme

Theme Resolving

You can enable a custom theme by creating a .vitepress/theme/index.js or .vitepress/theme/index.ts file (the "theme entry file"):

text
.
├─ docs                # project root
│  ├─ .vitepress
│  │  ├─ theme
│  │  │  └─ index.js   # theme entry
│  │  └─ config.js     # config file
│  └─ index.md
└─ package.json

VitePress will always use the custom theme instead of the default theme when it detects presence of a theme entry file. You can, however, extend the default theme to perform advanced customizations on top of it.

Theme Interface

A VitePress custom theme is defined as an object with the following interface:

ts
interface Theme {
  /**
   * Root layout component for every page
   * @required
   */
  Layout: Component
  /**
   * Enhance Vue app instance
   * @optional
   */
  enhanceApp?: (ctx: EnhanceAppContext) => Awaitable<void>
  /**
   * Extend another theme, calling its `enhanceApp` before ours
   * @optional
   */
  extends?: Theme
}

interface EnhanceAppContext {
  app: App // Vue app instance
  router: Router // VitePress router instance
  siteData: Ref<SiteData> // Site-level metadata
}

The theme entry file should export the theme as its default export:

js [.vitepress/theme/index.js]
// You can directly import Vue files in the theme entry
// VitePress is pre-configured with @vitejs/plugin-vue.
import Layout from './Layout.vue'

export default {
  Layout,
  enhanceApp({ app, router, siteData }) {
    // app.component(...)
    // app.use(...)
  }
}

The enhanceApp hook allows you to access the Vue app instance and other runtime data, which can be used to register global components, integrate with Vue libraries, etc.

The router value is the same VitePress router instance returned by useRouter(). To listen for route changes, assign handlers on the router:

ts [.vitepress/theme/index.ts]
export default {
  enhanceApp({ router }) {
    router.onBeforeRouteChange = (to) => {
      console.log('navigating to', to)
    }

    router.onAfterRouteChange = (to) => {
      console.log('navigated to', to)
    }
  }
}

Return false from onBeforeRouteChange or onBeforePageLoad to cancel navigation.

The default export is the only contract for a custom theme, and only the Layout property is required. So technically, a VitePress theme can be as simple as a single Vue component.

Inside your layout component, it works just like a normal Vite + Vue 3 application. Do note the theme also needs to be SSR-compatible.

Building a Layout

The most basic layout component needs to contain a <Content /> component:

vue [.vitepress/theme/layout.vue]
<template>
  <h1>Custom Layout!</h1>

  
  <Content />
</template>

The above layout simply renders every page's markdown as HTML. The first improvement we can add is to handle 404 errors:

vue{1-4,9-12}
<script setup>
import { useData } from 'vitepress'
const { page } = useData()
</script>

<template>
  <h1>Custom Layout!</h1>

  
    Custom 404 page!
  
  <Content v-else />
</template>

The useData() helper provides us with all the runtime data we need to conditionally render different layouts. One of the other data we can access is the current page's frontmatter. We can leverage this to allow the end user to control the layout in each page. For example, the user can indicate the page should use a special home page layout with:

md
---
layout: home
---

And we can adjust our theme to handle this:

vue{3,12-14}
<script setup>
import { useData } from 'vitepress'
const { page, frontmatter } = useData()
</script>

<template>
  <h1>Custom Layout!</h1>

  
    Custom 404 page!
  
  
    Custom home page!
  
  <Content v-else />
</template>

You can, of course, split the layout into more components:

vue{3-5,12-15}
<script setup>
import { useData } from 'vitepress'
import NotFound from './NotFound.vue'
import Home from './Home.vue'
import Page from './Page.vue'

const { page, frontmatter } = useData()
</script>

<template>
  <h1>Custom Layout!</h1>

  <NotFound v-if="page.isNotFound" />
  <Home v-if="frontmatter.layout === 'home'" />
   
</template>

Consult the Runtime API Reference for everything available in theme components. In addition, you can leverage Build-Time Data Loading to generate data-driven layout - for example, a page that lists all blog posts in the current project.

Distributing a Custom Theme

The easiest way to distribute a custom theme is by providing it as a template repository on GitHub.

If you wish to distribute the theme as an npm package, follow these steps:

  1. Export the theme object as the default export in your package entry.

  2. If applicable, export your theme config type definition as ThemeConfig.

  3. If your theme requires adjusting the VitePress config, export that config under a package sub-path (e.g. my-theme/config) so the user can extend it.

  4. Document the theme config options (both via config file and frontmatter).

  5. Provide clear instructions on how to consume your theme (see below).

Consuming a Custom Theme

To consume an external theme, import and re-export it from the custom theme entry:

js [.vitepress/theme/index.js]
import Theme from 'awesome-vitepress-theme'

export default Theme

If the theme needs to be extended:

js [.vitepress/theme/index.js]
import Theme from 'awesome-vitepress-theme'

export default {
  extends: Theme,
  enhanceApp(ctx) {
    // ...
  }
}

If the theme requires special VitePress config, you will need to also extend it in your own config:

ts [.vitepress/config.ts]
import baseConfig from 'awesome-vitepress-theme/config'

export default {
  // extend theme base config (if needed)
  extends: baseConfig
}

Finally, if the theme provides types for its theme config:

ts [.vitepress/config.ts]
import baseConfig from 'awesome-vitepress-theme/config'
import { defineConfig } from 'vitepress'
import type { ThemeConfig } from 'awesome-vitepress-theme'

export default defineConfig({
  extends: baseConfig,
  themeConfig: {
    // Type is `ThemeConfig`
  }
})

File: docs/en/guide/data-loading.md



Build-Time Data Loading

VitePress provides a feature called data loaders that allows you to load arbitrary data and import it from pages or components. The data loading is executed only at build time: the resulting data will be serialized as JSON in the final JavaScript bundle.

Data loaders can be used to fetch remote data, or generate metadata based on local files. For example, you can use data loaders to parse all your local API pages and automatically generate an index of all API entries.

Basic Usage

A data loader file must end with either .data.js or .data.ts. The file should provide a default export of an object with the load() method:

js [example.data.js]
export default {
  load() {
    return {
      hello: 'world'
    }
  }
}

The loader module is evaluated only in Node.js, so you can import Node APIs and npm dependencies as needed.

You can then import data from this file in .md pages and .vue components using the data named export:

vue
<script setup>
import { data } from './example.data.js'
</script>

{{ data }}

Output:

json
{
  "hello": "world"
}

You'll notice the data loader itself does not export the data. It is VitePress calling the load() method behind the scenes and implicitly exposing the result via the data named export.

This works even if the loader is async:

js
export default {
  async load() {
    // fetch remote data
    return (await fetch('...')).json()
  }
}

Data from Local Files

When you need to generate data based on local files, you should use the watch option in the data loader so that changes made to these files can trigger hot updates.

The watch option is also convenient in that you can use glob patterns to match multiple files. The patterns can be relative to the loader file itself, and the load() function will receive the matched files as absolute paths.

The following example shows loading CSV files and transforming them into JSON using csv-parse. Because this file only executes at build time, you will not be shipping the CSV parser to the client!

js
import fs from 'node:fs'
import { parse } from 'csv-parse/sync'

export default {
  watch: ['./data/*.csv'],
  load(watchedFiles) {
    // watchedFiles will be an array of absolute paths of the matched files.
    // generate an array of blog post metadata that can be used to render
    // a list in the theme layout
    return watchedFiles.map((file) => {
      return parse(fs.readFileSync(file, 'utf-8'), {
        columns: true,
        skip_empty_lines: true
      })
    })
  }
}

`createContentLoader`

When building a content focused site, we often need to create an "archive" or "index" page: a page where we list all available entries in our content collection, for example blog posts or API pages. We can implement this directly with the data loader API, but since this is such a common use case, VitePress also provides a createContentLoader helper to simplify this:

js [posts.data.js]
import { createContentLoader } from 'vitepress'

export default createContentLoader('posts/*.md', /* options */)

The helper takes a glob pattern relative to the source directory, and returns a { watch, load } data loader object that can be used as the default export in a data loader file. It also implements caching based on file modified timestamps to improve dev performance.

Note the loader only works with Markdown files - matched non-Markdown files will be skipped.

The loaded data will be an array with the type of ContentData[]:

ts
interface ContentData {
  // mapped URL for the page. e.g. /posts/hello.html (does not include base)
  // manually iterate or use custom `transform` to normalize the paths
  url: string
  // frontmatter data of the page
  frontmatter: Record<string, any>

  // the following are only present if relevant options are enabled
  // we will discuss them below
  src: string | undefined
  html: string | undefined
  excerpt: string | undefined
}

By default, only url and frontmatter are provided. This is because the loaded data will be inlined as JSON in the client bundle, so we need to be cautious about its size. Here's an example using the data to build a minimal blog index page:

vue
<script setup>
import { data as posts } from './posts.data.js'
</script>

<template>
  <h1>All Blog Posts</h1>
  
    <li v-for="post of posts">
      <a :href="post.url">{{ post.frontmatter.title }}</a>
      by {{ post.frontmatter.author }}
    </li>
  
</template>

Options

The default data may not suit all needs - you can opt-in to transform the data using options:

js [posts.data.js]
import { createContentLoader } from 'vitepress'

export default createContentLoader('posts/*.md', {
  includeSrc: true, // include raw markdown source?
  render: true,     // include rendered full page HTML?
  excerpt: true,    // include excerpt?
  transform(rawData) {
    // map, sort, or filter the raw data as you wish.
    // the final result is what will be shipped to the client.
    return rawData.sort((a, b) => {
      return +new Date(b.frontmatter.date) - +new Date(a.frontmatter.date)
    }).map((page) => {
      page.src     // raw markdown source
      page.html    // rendered full page HTML
      page.excerpt // rendered excerpt HTML (content above first `---`)
      return {/* ... */}
    })
  }
})

Check out how it is used in the Vue.js blog.

The createContentLoader API can also be used inside build hooks:

js [.vitepress/config.js]
export default {
  async buildEnd() {
    const posts = await createContentLoader('posts/*.md').load()
    // generate files based on posts metadata, e.g. RSS feed
  }
}

Types

ts
interface ContentOptions<T = ContentData[]> {
  /**
   * Include src?
   * @default false
   */
  includeSrc?: boolean

  /**
   * Render src to HTML and include in data?
   * @default false
   */
  render?: boolean

  /**
   * If `boolean`, whether to parse and include excerpt? (rendered as HTML)
   *
   * If `function`, control how the excerpt is extracted from the content.
   *
   * If `string`, define a custom separator to be used for extracting the
   * excerpt. Default separator is `---` if `excerpt` is `true`.
   *
   * @see https://github.com/jonschlinkert/gray-matter#optionsexcerpt
   * @see https://github.com/jonschlinkert/gray-matter#optionsexcerpt_separator
   *
   * @default false
   */
  excerpt?:
    | boolean
    | ((file: { data: { [key: string]: any }; content: string; excerpt?: string }, options?: any) => void)
    | string

  /**
   * Transform the data. Note the data will be inlined as JSON in the client
   * bundle if imported from components or markdown files.
   */
  transform?: (data: ContentData[]) => T | Promise<T>
}

Typed Data Loaders

When using TypeScript, you can type your loader and data export like so:

ts
import { defineLoader } from 'vitepress'

export interface Data {
  // data type
}

declare const data: Data
export { data }

export default defineLoader({
  // type checked loader options
  watch: ['...'],
  async load(): Promise<Data> {
    // ...
  }
})

Configuration

To get the configuration information inside a loader, you can use some code like this:

ts
import type { SiteConfig } from 'vitepress'

const config: SiteConfig = (globalThis as any).VITEPRESS_CONFIG

File: docs/en/guide/deploy.md


outline: deep

Deploy Your VitePress Site

The following guides are based on some shared assumptions:

  • The VitePress site is inside the docs directory of your project.

  • You are using the default build output directory (.vitepress/dist).

  • VitePress is installed as a local dependency in your project, and you have set up the following scripts in your package.json:

    json [package.json]
    {
      "scripts": {
        "docs:build": "vitepress build docs",
        "docs:preview": "vitepress preview docs"
      }
    }

Build and Test Locally

  1. Run this command to build the docs:

    sh
    $ npm run docs:build
  2. Once built, preview it locally by running:

    sh
    $ npm run docs:preview

    The preview command will boot up a local static web server that will serve the output directory .vitepress/dist at http://localhost:4173. You can use this to make sure everything looks good before pushing to production.

  3. You can configure the port of the server by passing --port as an argument.

    json
    {
      "scripts": {
        "docs:preview": "vitepress preview docs --port 8080"
      }
    }

    Now the docs:preview method will launch the server at http://localhost:8080.

Setting a Public Base Path

By default, we assume the site is going to be deployed at the root path of a domain (/). If your site is going to be served at a sub-path, e.g. https://mywebsite.com/blog/, then you need to set the base option to '/blog/' in the VitePress config.

Example: If you're using Github (or GitLab) Pages and deploying to user.github.io/repo/, then set your base to /repo/.

HTTP Cache Headers

If you have control over the HTTP headers on your production server, you can configure cache-control headers to achieve better performance on repeated visits.

The production build uses hashed file names for static assets (JavaScript, CSS and other imported assets not in public). If you inspect the production preview using your browser devtools' network tab, you will see files like app.4f283b18.js.

This 4f283b18 hash is generated from the content of this file. The same hashed URL is guaranteed to serve the same file content - if the contents change, the URLs change too. This means you can safely use the strongest cache headers for these files. All such files will be placed under assets/ in the output directory, so you can configure the following header for them:

text
Cache-Control: max-age=31536000,immutable

::: details Example Netlify _headers file

text
/assets/*
  cache-control: max-age=31536000
  cache-control: immutable

Note: the _headers file should be placed in the public directory - in our case, docs/public/_headers - so that it is copied verbatim to the output directory.

Netlify custom headers documentation

:::

::: details Example Vercel config in vercel.json

json
{
  "headers": [
    {
      "source": "/assets/(.*)",
      "headers": [
        {
          "key": "Cache-Control",
          "value": "max-age=31536000, immutable"
        }
      ]
    }
  ]
}

Note: the vercel.json file should be placed at the root of your repository.

Vercel documentation on headers config

:::

Platform Guides

Netlify / Vercel / Cloudflare Pages / AWS Amplify / Render {#generic}

Set up a new project and change these settings using your dashboard:

  • Build Command: npm run docs:build
  • Output Directory: docs/.vitepress/dist
  • Node Version: 20 (or above)

::: warning
Don't enable options like Auto Minify for HTML code. It will remove comments from output which have meaning to Vue. You may see hydration mismatch errors if they get removed.
:::

GitHub Pages

  1. Create a file named deploy.yml inside .github/workflows directory of your project with some content like this:

    text

/* Detailed source-code truncated for AI context efficiency. */

text
::: warning
   Make sure the `base` option in your VitePress is properly configured. See [Setting a Public Base Path](#setting-a-public-base-path) for more details.
   :::

2. In your repository's settings under "Pages" menu item, select "GitHub Actions" in "Build and deployment > Source".

3. Push your changes to the `main` branch and wait for the GitHub Actions workflow to complete. You should see your site deployed to `https://.github.io/[repository]/` or `https://<custom-domain>/` depending on your settings. Your site will automatically be deployed on every push to the `main` branch.

### GitLab Pages

1. Set `outDir` in VitePress config to `../public`. Configure `base` option to `'/<repository>/'` if you want to deploy to `https://.gitlab.io/<repository>/`. You don't need `base` if you're deploying to custom domain, user or group pages, or have "Use unique domain" setting enabled in GitLab.

2. Create a file named `.gitlab-ci.yml` in the root of your project with the content below. This will build and deploy your site whenever you make changes to your content:

   ```yaml [.gitlab-ci.yml]
   image: node:24
   pages:
     cache:
       paths:
         - node_modules/
     script:
       # - apk add git # Uncomment this if you're using small docker images like alpine and have lastUpdated enabled
       - npm install
       - npm run docs:build
     artifacts:
       paths:
         - public
     only:
       - main

Azure

  1. Follow the official documentation.

  2. Set these values in your configuration file (and remove the ones you don't require, like api_location):

    • app_location: /
    • output_location: docs/.vitepress/dist
    • app_build_command: npm run docs:build

CloudRay

You can deploy your VitePress project with CloudRay by following these instructions.

Firebase

  1. Create firebase.json and .firebaserc at the root of your project:

    firebase.json:

    json [firebase.json]
    {
      "hosting": {
        "public": "docs/.vitepress/dist",
        "ignore": []
      }
    }

    .firebaserc:

    json [.firebaserc]
    {
      "projects": {
        "default": "<YOUR_FIREBASE_ID>"
      }
    }
  2. After running npm run docs:build, run this command to deploy:

    sh
    firebase deploy

Heroku

  1. Follow documentation and guide given in heroku-buildpack-static.

  2. Create a file called static.json in the root of your project with the below content:

    json [static.json]
    {
      "root": "docs/.vitepress/dist"
    }

Hostinger

You can deploy your VitePress project with Hostinger by following these instructions. While configuring build settings, choose VitePress as the framework and adjust the root directory to ./docs.

Stormkit

You can deploy your VitePress project to Stormkit by following these instructions.

Surge

After running npm run docs:build, run this command to deploy to Surge:

sh
npx surge docs/.vitepress/dist

harvis

After running npm run docs:build, run this command to deploy to harvis:

sh
npx harvis docs/.vitepress/dist

nginx

Here is a example of an nginx server block configuration. This setup includes gzip compression for common text-based assets, rules for serving your VitePress site's static files with proper caching headers as well as handling cleanUrls: true.

nginx
map $uri $cache_control {
    ~^/assets/  "public, max-age=31536000, immutable";
    default     "no-cache";
}

server {
    listen 8080;
    listen [::]:8080;
    server_name _;

    root /usr/share/nginx/html;
    index index.html;
    charset utf-8;
    server_tokens off;

    absolute_redirect off;

    gzip on;
    gzip_vary on;
    gzip_comp_level 5;
    gzip_min_length 1024;
    gzip_types
        application/javascript
        application/json
        application/manifest+json
        image/svg+xml
        text/css
        text/javascript
        text/plain;

    add_header Cache-Control $cache_control always;
    add_header X-Content-Type-Options "nosniff" always;
    add_header X-Frame-Options "SAMEORIGIN" always;
    add_header Referrer-Policy "strict-origin-when-cross-origin" always;

    location / {
        try_files $uri $uri.html $uri/index.html =404;
    }

    location ~ ^(?.+)/$ {
        if (-f $document_root$page.html) {
            return 301 $page$is_args$args;
        }
        try_files $page/index.html =404;
    }

    error_page 404 /404.html;
}

File: docs/en/guide/extending-default-theme.md


outline: deep

Extending the Default Theme

VitePress' default theme is optimized for documentation, and can be customized. Consult the Default Theme Config Overview for a comprehensive list of options.

However, there are a number of cases where configuration alone won't be enough. For example:

  1. You need to tweak the CSS styling;
  2. You need to modify the Vue app instance, for example to register global components;
  3. You need to inject custom content into the theme via layout slots.

These advanced customizations will require using a custom theme that "extends" the default theme.

::: tip
Before proceeding, make sure to first read Using a Custom Theme to understand how custom themes work.
:::

Customizing CSS

The default theme CSS is customizable by overriding root level CSS variables:

js [.vitepress/theme/index.js]
import DefaultTheme from 'vitepress/theme'
import './custom.css'

export default DefaultTheme
css
/* .vitepress/theme/custom.css */
:root {
  --vp-c-brand-1: #646cff;
  --vp-c-brand-2: #747bff;
}

See default theme CSS variables that can be overridden.

Using Different Fonts

VitePress uses Inter as the default font, and will include the fonts in the build output. The font is also auto preloaded in production. However, this may not be desirable if you want to use a different main font.

To avoid including Inter in the build output, import the theme from vitepress/theme-without-fonts instead:

js [.vitepress/theme/index.js]
import DefaultTheme from 'vitepress/theme-without-fonts'
import './my-fonts.css'

export default DefaultTheme
css
/* .vitepress/theme/my-fonts.css */
:root {
  --vp-font-family-base: /* normal text font */
  --vp-font-family-mono: /* code font */
}

::: warning
If you are using optional components like the Team Page components, make sure to also import them from vitepress/theme-without-fonts!
:::

If your font is a local file referenced via @font-face, it will be processed as an asset and included under .vitepress/dist/assets with hashed filename. To preload this file, use the transformHead build hook:

js [.vitepress/config.js]
export default {
  transformHead({ assets }) {
    // adjust the regex accordingly to match your font
    const myFontFile = assets.find(file => /font-name\.[\w-]+\.woff2/.test(file))
    if (myFontFile) {
      return [
        [
          'link',
          {
            rel: 'preload',
            href: myFontFile,
            as: 'font',
            type: 'font/woff2',
            crossorigin: ''
          }
        ]
      ]
    }
  }
}

Registering Global Components

js [.vitepress/theme/index.js]
import DefaultTheme from 'vitepress/theme'

/** @type {import('vitepress').Theme} */
export default {
  extends: DefaultTheme,
  enhanceApp({ app }) {
    // register your custom global components
    app.component('MyGlobalComponent' /* ... */)
  }
}

If you're using TypeScript:

ts [.vitepress/theme/index.ts]
import type { Theme } from 'vitepress'
import DefaultTheme from 'vitepress/theme'

export default {
  extends: DefaultTheme,
  enhanceApp({ app }) {
    // register your custom global components
    app.component('MyGlobalComponent' /* ... */)
  }
} satisfies Theme

Since we are using Vite, you can also leverage Vite's glob import feature to auto register a directory of components.

Layout Slots

The default theme's <Layout/> component has a few slots that can be used to inject content at certain locations of the page. Here's an example of injecting a component into the before outline:

js [.vitepress/theme/index.js]
import DefaultTheme from 'vitepress/theme'
import MyLayout from './MyLayout.vue'

export default {
  extends: DefaultTheme,
  // override the Layout with a wrapper component that
  // injects the slots
  Layout: MyLayout
}
vue [.vitepress/theme/mylayout.vue]
<script setup>
import DefaultTheme from 'vitepress/theme'

const { Layout } = DefaultTheme
</script>

<template>
  <Layout>
    <template #aside-outline-before>
      My custom sidebar top content
    </template>
  </Layout>
</template>

Or you could use render function as well.

js [.vitepress/theme/index.js]
import { h } from 'vue'
import DefaultTheme from 'vitepress/theme'
import MyComponent from './MyComponent.vue'

export default {
  extends: DefaultTheme,
  Layout() {
    return h(DefaultTheme.Layout, null, {
      'aside-outline-before': () => h(MyComponent)
    })
  }
}

Full list of slots available in the default theme layout:

  • When layout: 'doc' (default) is enabled via frontmatter:
    • doc-top
    • doc-bottom
    • doc-footer-before
    • doc-before
    • doc-after
    • sidebar-nav-before
    • sidebar-nav-after
    • aside-top
    • aside-bottom
    • aside-outline-before
    • aside-outline-after
    • aside-ads-before
    • aside-ads-after
  • When layout: 'home' is enabled via frontmatter:
    • home-hero-before
    • home-hero-info-before
    • home-hero-info
    • home-hero-info-after
    • home-hero-actions-before-actions
    • home-hero-actions-after
    • home-hero-image
    • home-hero-after
    • home-features-before
    • home-features-after
  • When layout: 'page' is enabled via frontmatter:
    • page-top
    • page-bottom
  • On not found (404) page:
    • not-found
  • Always:
    • layout-top
    • layout-bottom
    • nav-bar-title-before
    • nav-bar-title-after
    • nav-bar-content-before
    • nav-bar-content-after
    • nav-screen-content-before
    • nav-screen-content-after

Using View Transitions API

On Appearance Toggle

You can extend the default theme to provide a custom transition when the color mode is toggled. An example:

vue [.vitepress/theme/layout.vue]
<script setup lang="ts">
import { useData } from 'vitepress'
import DefaultTheme from 'vitepress/theme'
import { nextTick, provide } from 'vue'

const { isDark } = useData()

const enableTransitions = () =>
  'startViewTransition' in document &&
  window.matchMedia('(prefers-reduced-motion: no-preference)').matches

provide('toggle-appearance', async ({ clientX: x, clientY: y }: MouseEvent) => {
  if (!enableTransitions()) {
    isDark.value = !isDark.value
    return
  }

  const clipPath = [
    `circle(0px at ${x}px ${y}px)`,
    `circle(${Math.hypot(
      Math.max(x, innerWidth - x),
      Math.max(y, innerHeight - y)
    )}px at ${x}px ${y}px)`
  ]

  await document.startViewTransition(async () => {
    isDark.value = !isDark.value
    await nextTick()
  }).ready

  document.documentElement.animate(
    { clipPath: isDark.value ? clipPath.reverse() : clipPath },
    {
      duration: 300,
      easing: 'ease-in',
      fill: 'forwards',
      pseudoElement: `::view-transition-${isDark.value ? 'old' : 'new'}(root)`
    }
  )
})
</script>

<template>
  <DefaultTheme.Layout />
</template>

<style>
::view-transition-old(root),
::view-transition-new(root) {
  animation: none;
  mix-blend-mode: normal;
}

::view-transition-old(root),
.dark::view-transition-new(root) {
  z-index: 1;
}

::view-transition-new(root),
.dark::view-transition-old(root) {
  z-index: 9999;
}

.VPSwitchAppearance {
  width: 22px !important;
}

.VPSwitchAppearance .check {
  transform: none !important;
}
</style>

Result (warning!: flashing colors, sudden movements, bright lights):

Demo

Refer Chrome Docs from more details on view transitions.

On Route Change

Coming soon.

Overriding Internal Components

You can use Vite's aliases to replace default theme components with your custom ones:

ts
import { fileURLToPath, URL } from 'node:url'
import { defineConfig } from 'vitepress'

export default defineConfig({
  vite: {
    resolve: {
      alias: [
        {
          find: /^.*\/VPNavBar\.vue$/,
          replacement: fileURLToPath(
            new URL('./theme/components/CustomNavBar.vue', import.meta.url)
          )
        }
      ]
    }
  }
})

To know the exact name of the component refer our source code. Since the components are internal, there is a slight chance their name is updated between minor releases.


File: docs/en/guide/frontmatter.md



Frontmatter

Usage

VitePress supports YAML frontmatter in all Markdown files, parsing them with gray-matter. The frontmatter must be at the top of the Markdown file (before any elements including <script> tags), and must take the form of valid YAML set between triple-dashed lines. Example:

md
---
title: Docs with VitePress
editLink: true
---

Many site or default theme config options have corresponding options in frontmatter. You can use frontmatter to override specific behavior for the current page only. For details, see Frontmatter Config Reference.

You can also define custom frontmatter data of your own, to be used in dynamic Vue expressions on the page.

Accessing Frontmatter Data

Frontmatter data can be accessed via the special $frontmatter global variable:

Here's an example of how you could use it in your Markdown file:

md
---
title: Docs with VitePress
editLink: true
---

# {{ $frontmatter.title }}

Guide content

You can also access current page's frontmatter data in <script setup> with the useData() helper.

Alternative Frontmatter Formats

VitePress also supports JSON frontmatter syntax, starting and ending in curly braces:

json
---
{
  "title": "Blogging Like a Hacker",
  "editLink": true
}
---

File: docs/en/guide/getting-started.md



Getting Started

Try It Online

You can try VitePress directly in your browser on StackBlitz.

Installation

Prerequisites

VitePress can be used on its own, or be installed into an existing project. In both cases, you can install it with:

::: code-group

sh [npm]
$ npm add -D vitepress@next
sh [pnpm]
$ pnpm add -D vitepress@next
sh [yarn]
$ yarn add -D vitepress@next vue
sh [bun]
$ bun add -D vitepress@next
sh [deno]
$ deno add -D vitepress@next

:::

::: tip NOTE

VitePress is an ESM-only package. Don't use require() to import it, and make sure your nearest package.json contains "type": "module", or change the file extension of your relevant files like .vitepress/config.js to .mjs/.mts. Refer to Vite's troubleshooting guide for more details. Also, inside async CJS contexts, you can use await import('vitepress') instead.

:::

Setup Wizard

VitePress ships with a command line setup wizard that will help you scaffold a basic project. After installation, start the wizard by running:

::: code-group

sh [npm]
$ npx vitepress init
sh [pnpm]
$ pnpm vitepress init
sh [yarn]
$ yarn vitepress init
sh [bun]
$ bun vitepress init

:::

You will be greeted with a few simple questions:

<<< @/snippets/init.ansi

::: tip Vue as Peer Dependency
If you intend to perform customization that uses Vue components or APIs, you should also explicitly install vue as a dependency.
:::

File Structure

If you are building a standalone VitePress site, you can scaffold the site in your current directory (./). However, if you are installing VitePress in an existing project alongside other source code, it is recommended to scaffold the site in a nested directory (e.g. ./docs) so that it is separate from the rest of the project.

Assuming you chose to scaffold the VitePress project in ./docs, the generated file structure should look like this:

text
.
├─ docs
│  ├─ .vitepress
│  │  └─ config.js
│  ├─ api-examples.md
│  ├─ markdown-examples.md
│  └─ index.md
└─ package.json

The docs directory is considered the project root of the VitePress site. The .vitepress directory is a reserved location for VitePress' config file, dev server cache, build output, and optional theme customization code.

::: tip
By default, VitePress stores its dev server cache in .vitepress/cache, and the production build output in .vitepress/dist. If using Git, you should add them to your .gitignore file. These locations can also be configured.
:::

The Config File

The config file (.vitepress/config.js) allows you to customize various aspects of your VitePress site, with the most basic options being the title and description of the site:

js [.vitepress/config.js]
export default {
  // site-level options
  title: 'VitePress',
  themeConfig: {
    // theme-level options
  }
}

You can also configure the behavior of the theme via the themeConfig option. Consult the Config Reference for full details on all config options.

Source Files

Markdown files outside the .vitepress directory are considered source files.

VitePress uses file-based routing: each .md file is compiled into a corresponding .html file with the same path. For example, index.md will be compiled into index.html, and can be visited at the root path / of the resulting VitePress site.

VitePress also provides the ability to generate clean URLs, rewrite paths, and dynamically generate pages. These will be covered in the Routing Guide.

Up and Running

The tool should have also injected the following npm scripts to your package.json if you allowed it to do so during the setup process:

json [package.json]
{
  ...
  "scripts": {
    "docs:dev": "vitepress dev docs",
    "docs:build": "vitepress build docs",
    "docs:preview": "vitepress preview docs"
  },
  ...
}

The docs:dev script will start a local dev server with instant hot updates. Run it with the following command:

::: code-group

sh [npm]
$ npm run docs:dev
sh [pnpm]
$ pnpm run docs:dev
sh [yarn]
$ yarn docs:dev
sh [bun]
$ bun run docs:dev

:::

Instead of npm scripts, you can also invoke VitePress directly with:

::: code-group

sh [npm]
$ npx vitepress dev docs
sh [pnpm]
$ pnpm vitepress dev docs
sh [yarn]
$ yarn vitepress dev docs
sh [bun]
$ bun vitepress dev docs

:::

More command line usage is documented in the CLI Reference.

The dev server should be running at http://localhost:5173. Visit the URL in your browser to see your new site in action!

What's Next?

  • To better understand how markdown files are mapped to generated HTML, proceed to the Routing Guide.

  • To discover more about what you can do on the page, such as writing markdown content or using Vue Components, refer to the "Writing" section of the guide. A great place to start would be to learn about Markdown Extensions.

  • To explore the features provided by the default documentation theme, check out the Default Theme Config Reference.

  • If you want to further customize the appearance of your site, explore how to either Extend the Default Theme or Build a Custom Theme.

  • Once your documentation site takes shape, make sure to read the Deployment Guide.


File: docs/en/guide/i18n.md



Internationalization

To use the built-in i18n features, one needs to create a directory structure as follows:

text
docs/
├─ es/
│  ├─ foo.md
├─ fr/
│  ├─ foo.md
├─ foo.md

Then in docs/.vitepress/config.ts:

ts [docs/.vitepress/config.ts]
import { defineConfig } from 'vitepress'

export default defineConfig({
  // shared properties and other top-level stuff...

  locales: {
    root: {
      label: 'English',
      lang: 'en'
    },
    fr: {
      label: 'French',
      lang: 'fr', // optional, will be added  as `lang` attribute on `html` tag
      link: '/fr/guide' // default /fr/ -- shows on navbar translations menu, can be external

      // other locale specific properties...
    }
  }
})

The following properties can be overridden for each locale (including root):

ts
interface LocaleSpecificConfig {
  lang?: string
  dir?: string
  title?: string
  titleTemplate?: string | boolean
  description?: string
  head?: HeadConfig[] // will be merged with existing head entries, duplicate meta tags are automatically removed
  themeConfig?: ThemeConfig // will be shallow merged, common stuff can be put in top-level themeConfig entry
}

Refer DefaultTheme.Config interface for details on customizing the placeholder texts of the default theme. Don't override themeConfig.algolia or themeConfig.carbonAds at locale-level. Refer Algolia docs for using multilingual search.

Pro tip: Config file can be stored at docs/.vitepress/config/index.ts too. It might help you organize stuff by creating a configuration file per locale and then merge and export them from index.ts.

Per-locale Markdown Strings

Strings baked into pages by the markdown renderer - the default titles of custom containers and GitHub-flavored alerts, and the strings of the code copy button - can be overridden per locale with the markdown key of a locale entry:

ts [docs/.vitepress/config.ts]
import { defineConfig } from 'vitepress'

export default defineConfig({
  locales: {
    root: { label: 'English', lang: 'en' },
    zh: {
      label: '简体中文',
      lang: 'zh-Hans',
      markdown: {
        container: {
          tipLabel: '提示',
          warningLabel: '警告'
          // ...the other labels, and titles of `customContainers`
        },
        codeCopyButton: {
          tooltipText: '复制代码',
          copiedText: '已复制'
        }
      }
    }
  }
})

Values fall back to the root-level markdown options when a locale leaves them unset. Locale entries can only override the titles of containers registered at the root level - registering new containers per locale is not supported. Also note that since the markdown renderer is created once for the whole site, these can only be declared in the main config file, not in additional configs.

Separate directory for each locale

The following is a perfectly fine structure:

text
docs/
├─ en/
│  ├─ foo.md
├─ es/
│  ├─ foo.md
├─ fr/
   ├─ foo.md

However, VitePress won't redirect / to /en/ by default. You'll need to configure your server for that. For example, on Netlify, you can add a docs/public/_redirects file like this:

text
/*  /es/:splat  302  Language=es
/*  /fr/:splat  302  Language=fr
/*  /en/:splat  302

Pro tip: If using the above approach, you can use nf_lang cookie to persist user's language choice:

ts [docs/.vitepress/theme/index.ts]
import DefaultTheme from 'vitepress/theme'
import Layout from './Layout.vue'

export default {
  extends: DefaultTheme,
  Layout
}
vue [docs/.vitepress/theme/layout.vue]
<script setup lang="ts">
import DefaultTheme from 'vitepress/theme'
import { useData, inBrowser } from 'vitepress'
import { watchEffect } from 'vue'

const { lang } = useData()
watchEffect(() => {
  if (inBrowser) {
    document.cookie = `nf_lang=${lang.value}; expires=Mon, 1 Jan 2030 00:00:00 UTC; path=/`
  }
})
</script>

<template>
  <DefaultTheme.Layout />
</template>

RTL Support (Experimental)

For RTL support, specify dir: 'rtl' in config and use some RTLCSS PostCSS plugin like https://github.com/MohammadYounes/rtlcss, https://github.com/vkalinichev/postcss-rtl or https://github.com/elchininet/postcss-rtlcss. You'll need to configure your PostCSS plugin to use :where([dir="ltr"]) and :where([dir="rtl"]) as prefixes to prevent CSS specificity issues.

2. Official Technical Reference & Guides (vuejs/docs)

File: README.md

vuejs.org

Contributing

This site is built with VitePress and depends on @vue/theme. Site content is written in Markdown format located in src. For simple edits, you can directly edit the file on GitHub and generate a Pull Request.

For local development, pnpm is preferred as package manager:

bash
pnpm i
pnpm run dev

This project requires Node.js to be v20 or higher. And it is recommended to enable corepack:

bash
corepack enable

Working on the content

Working on the theme

If changes need to be made for the theme, check out the instructions for developing the theme alongside the docs.


File: src/guide/extras/animation.md

Animation Techniques {#animation-techniques}

Vue provides the ](/guide/built-ins/transition) and [ components for handling enter / leave and list transitions. However, there are many other ways of using animations on the web, even in a Vue application. Here we will discuss a few additional techniques.

Class-based Animations {#class-based-animations}

For elements that are not entering / leaving the DOM, we can trigger animations by dynamically adding a CSS class:

js
const disabled = ref(false)

function warnDisabled() {
  disabled.value = true
  setTimeout(() => {
    disabled.value = false
  }, 1500)
}
js
export default {
  data() {
    return {
      disabled: false
    }
  },
  methods: {
    warnDisabled() {
      this.disabled = true
      setTimeout(() => {
        this.disabled = false
      }, 1500)
    }
  }
}
vue-html
Click me
  This feature is disabled!
css
.shake {
  animation: shake 0.82s cubic-bezier(0.36, 0.07, 0.19, 0.97) both;
  transform: translate3d(0, 0, 0);
}

@keyframes shake {
  10%,
  90% {
    transform: translate3d(-1px, 0, 0);
  }

  20%,
  80% {
    transform: translate3d(2px, 0, 0);
  }

  30%,
  50%,
  70% {
    transform: translate3d(-4px, 0, 0);
  }

  40%,
  60% {
    transform: translate3d(4px, 0, 0);
  }
}

State-driven Animations {#state-driven-animations}

Some transition effects can be applied by interpolating values, for instance by binding a style to an element while an interaction occurs. Take this example for instance:

js
const x = ref(0)

function onMousemove(e) {
  x.value = e.clientX
}
js
export default {
  data() {
    return {
      x: 0
    }
  },
  methods: {
    onMousemove(e) {
      this.x = e.clientX
    }
  }
}
vue-html
Move your mouse across this div...
  x: {{ x }}
css
.movearea {
  transition: 0.3s background-color ease;
}

In addition to color, you can also use style bindings to animate transform, width, or height. You can even animate SVG paths using spring physics - after all, they are all attribute data bindings:

Animating with Watchers {#animating-with-watchers}

With some creativity, we can use watchers to animate anything based on some numerical state. For example, we can animate the number itself:

js
import { ref, reactive, watch } from 'vue'
import gsap from 'gsap'

const number = ref(0)
const tweened = reactive({
  number: 0
})

// Note: For inputs greater than Number.MAX_SAFE_INTEGER (9007199254740991),
// the result may be inaccurate due to limitations in JavaScript number precision.
watch(number, (n) => {
  gsap.to(tweened, { duration: 0.5, number: Number(n) || 0 })
})
vue-html
Type a number: 
{{ tweened.number.toFixed(0) }}
js
import gsap from 'gsap'

export default {
  data() {
    return {
      number: 0,
      tweened: 0
    }
  },
  // Note: For inputs greater than Number.MAX_SAFE_INTEGER (9007199254740991),
  // the result may be inaccurate due to limitations in JavaScript number precision.
  watch: {
    number(n) {
      gsap.to(this, { duration: 0.5, tweened: Number(n) || 0 })
    }
  }
}
vue-html
Type a number: 
{{ tweened.toFixed(0) }}

Try it in the Playground

Try it in the Playground


File: src/guide/extras/composition-api-faq.md


outline: deep

Composition API FAQ {#composition-api-faq}

:::tip
This FAQ assumes prior experience with Vue - in particular, experience with Vue 2 while primarily using Options API.
:::

What is Composition API? {#what-is-composition-api}

Composition API is a set of APIs that allows us to author Vue components using imported functions instead of declaring options. It is an umbrella term that covers the following APIs:

  • Reactivity API, e.g. ref() and reactive(), that allows us to directly create reactive state, computed state, and watchers.

  • Lifecycle Hooks, e.g. onMounted() and onUnmounted(), that allow us to programmatically hook into the component lifecycle.

  • Dependency Injection, i.e. provide() and inject(), that allow us to leverage Vue's dependency injection system while using Reactivity APIs.

Composition API is a built-in feature of Vue 3 and Vue 2.7. For older Vue 2 versions, use the officially maintained @vue/composition-api plugin. In Vue 3, it is also primarily used together with the <script setup> syntax in Single-File Components. Here's a basic example of a component using Composition API:

vue
<script setup>
import { ref, onMounted } from 'vue'

// reactive state
const count = ref(0)

// functions that mutate state and trigger updates
function increment() {
  count.value++
}

// lifecycle hooks
onMounted(() => {
  console.log(`The initial count is ${count.value}.`)
})
</script>

<template>
  Count is: {{ count }}
</template>

Despite an API style based on function composition, Composition API is NOT functional programming. Composition API is based on Vue's mutable, fine-grained reactivity paradigm, whereas functional programming emphasizes immutability.

If you are interested in learning how to use Vue with Composition API, you can set the site-wide API preference to Composition API using the toggle at the top of the left sidebar, and then go through the guide from the beginning.

Why Composition API? {#why-composition-api}

Better Logic Reuse {#better-logic-reuse}

The primary advantage of Composition API is that it enables clean, efficient logic reuse in the form of Composable functions. It solves all the drawbacks of mixins, the primary logic reuse mechanism for Options API.

Composition API's logic reuse capability has given rise to impressive community projects such as VueUse, an ever-growing collection of composable utilities. It also serves as a clean mechanism for easily integrating stateful third-party services or libraries into Vue's reactivity system, for example immutable data, state machines, and RxJS.

More Flexible Code Organization {#more-flexible-code-organization}

Many users love that we write organized code by default with Options API: everything has its place based on the option it falls under. However, Options API poses serious limitations when a single component's logic grows beyond a certain complexity threshold. This limitation is particularly prominent in components that need to deal with multiple logical concerns, which we have witnessed first hand in many production Vue 2 apps.

Take the folder explorer component from Vue CLI's GUI as an example: this component is responsible for the following logical concerns:

  • Tracking current folder state and displaying its content
  • Handling folder navigation (opening, closing, refreshing...)
  • Handling new folder creation
  • Toggling show favorite folders only
  • Toggling show hidden folders
  • Handling current working directory changes

The original version of the component was written in Options API. If we give each line of code a color based on the logical concern it is dealing with, this is how it looks:

Notice how code dealing with the same logical concern is forced to be split under different options, located in different parts of the file. In a component that is several hundred lines long, understanding and navigating a single logical concern requires constantly scrolling up and down the file, making it much more difficult than it should be. In addition, if we ever intend to extract a logical concern into a reusable utility, it takes quite a bit of work to find and extract the right pieces of code from different parts of the file.

Here's the same component, before and after the refactor into Composition API:

Notice how the code related to the same logical concern can now be grouped together: we no longer need to jump between different options blocks while working on a specific logical concern. Moreover, we can now move a group of code into an external file with minimal effort, since we no longer need to shuffle the code around in order to extract them. This reduced friction for refactoring is key to the long-term maintainability in large codebases.

Better Type Inference {#better-type-inference}

In recent years, more and more frontend developers are adopting TypeScript as it helps us write more robust code, make changes with more confidence, and provides a great development experience with IDE support. However, the Options API, originally conceived in 2013, was designed without type inference in mind. We had to implement some absurdly complex type gymnastics to make type inference work with the Options API. Even with all this effort, type inference for Options API can still break down for mixins and dependency injection.

This had led many developers who wanted to use Vue with TS to lean towards Class API powered by vue-class-component. However, a class-based API heavily relies on ES decorators, a language feature that was only a stage 2 proposal when Vue 3 was being developed in 2019. We felt it was too risky to base an official API on an unstable proposal. Since then, the decorators proposal has gone through yet another complete overhaul, and finally reached stage 3 in 2022. In addition, class-based API suffers from logic reuse and organization limitations similar to Options API.

In comparison, Composition API utilizes mostly plain variables and functions, which are naturally type friendly. Code written in Composition API can enjoy full type inference with little need for manual type hints. Most of the time, Composition API code will look largely identical in TypeScript and plain JavaScript. This also makes it possible for plain JavaScript users to benefit from partial type inference.

Smaller Production Bundle and Less Overhead {#smaller-production-bundle-and-less-overhead}

Code written in Composition API and <script setup> is also more efficient and minification-friendly than Options API equivalent. This is because the template in a <script setup> component is compiled as a function inlined in the same scope of the <script setup> code. Unlike property access from this, the compiled template code can directly access variables declared inside <script setup>, without an instance proxy in between. This also leads to better minification because all the variable names can be safely shortened.

Relationship with Options API {#relationship-with-options-api}

Trade-offs {#trade-offs}

Some users moving from Options API found their Composition API code less organized, and concluded that Composition API is "worse" in terms of code organization. We recommend users with such opinions to look at that problem from a different perspective.

It is true that Composition API no longer provides the "guard rails" that guide you to put your code into respective buckets. In return, you get to author component code like how you would write normal JavaScript. This means you can and should apply any code organization best practices to your Composition API code as you would when writing normal JavaScript. If you can write well-organized JavaScript, you should also be able to write well-organized Composition API code.

Options API does allow you to "think less" when writing component code, which is why many users love it. However, in reducing the mental overhead, it also locks you into the prescribed code organization pattern with no escape hatch, which can make it difficult to refactor or improve code quality in larger scale projects. In this regard, Composition API provides better long term scalability.

Does Composition API cover all use cases? {#does-composition-api-cover-all-use-cases}

Yes in terms of stateful logic. When using Composition API, there are only a few options that may still be needed: props, emits, name, and inheritAttrs.

:::tip

Since 3.3 you can directly use defineOptions in <script setup> to set the component name or inheritAttrs property

:::

If you intend to exclusively use Composition API (along with the options listed above), you can shave a few kbs off your production bundle via a compile-time flag that drops Options API related code from Vue. Note this also affects Vue components in your dependencies.

Can I use both APIs in the same component? {#can-i-use-both-apis-in-the-same-component}

Yes. You can use Composition API via the setup() option in an Options API component.

However, we only recommend doing so if you have an existing Options API codebase that needs to integrate with new features / external libraries written with Composition API.

Will Options API be deprecated? {#will-options-api-be-deprecated}

No, we do not have any plan to do so. Options API is an integral part of Vue and the reason many developers love it. We also realize that many of the benefits of Composition API only manifest in larger-scale projects, and Options API remains a solid choice for many low-to-medium-complexity scenarios.

Relationship with Class API {#relationship-with-class-api}

We no longer recommend using Class API with Vue 3, given that Composition API provides great TypeScript integration with additional logic reuse and code organization benefits.

Comparison with React Hooks {#comparison-with-react-hooks}

Composition API provides the same level of logic composition capabilities as React Hooks, but with some important differences.

React Hooks are invoked repeatedly every time a component updates. This creates a number of caveats that can confuse even seasoned React developers. It also leads to performance optimization issues that can severely affect development experience. Here are some examples:

  • Hooks are call-order sensitive and cannot be conditional.

  • Variables declared in a React component can be captured by a hook closure and become "stale" if the developer fails to pass in the correct dependencies array. This leads to React developers relying on ESLint rules to ensure correct dependencies are passed. However, the rule is often not smart enough and over-compensates for correctness, which leads to unnecessary invalidation and headaches when edge cases are encountered.

  • Expensive computations require the use of useMemo, which again requires manually passing in the correct dependencies array.

  • Event handlers passed to child components cause unnecessary child updates by default, and require explicit useCallback as an optimization. This is almost always needed, and again requires a correct dependencies array. Neglecting this leads to over-rendering apps by default and can cause performance issues without realizing it.

  • The stale closure problem, combined with Concurrent features, makes it difficult to reason about when a piece of hooks code is run, and makes working with mutable state that should persist across renders (via useRef) cumbersome.

Note: some of the above issues that are related to memoization can be resolved by the upcoming React Compiler.

In comparison, Vue Composition API:

  • Invokes setup() or <script setup> code only once. This makes the code align better with the intuitions of idiomatic JavaScript usage as there are no stale closures to worry about. Composition API calls are also not sensitive to call order and can be conditional.

  • Vue's runtime reactivity system automatically collects reactive dependencies used in computed properties and watchers, so there's no need to manually declare dependencies.

  • No need to manually cache callback functions to avoid unnecessary child updates. In general, Vue's fine-grained reactivity system ensures child components only update when they need to. Manual child-update optimizations are rarely a concern for Vue developers.

We acknowledge the creativity of React Hooks, and it is a major source of inspiration for Composition API. However, the issues mentioned above do exist in its design and we noticed Vue's reactivity model happens to provide a way around them.


File: src/guide/extras/reactivity-in-depth.md


outline: deep

Reactivity in Depth {#reactivity-in-depth}

One of Vue’s most distinctive features is the unobtrusive reactivity system. Component state consists of reactive JavaScript objects. When you modify them, the view updates. It makes state management simple and intuitive, but it’s also important to understand how it works to avoid some common gotchas. In this section, we are going to dig into some of the lower-level details of Vue’s reactivity system.

What is Reactivity? {#what-is-reactivity}

This term comes up in programming quite a bit these days, but what do people mean when they say it? Reactivity is a programming paradigm that allows us to adjust to changes in a declarative manner. The canonical example that people usually show, because it’s a great one, is an Excel spreadsheet:

Here cell A2 is defined via a formula of = A0 + A1 (you can click on A2 to view or edit the formula), so the spreadsheet gives us 3. No surprises there. But if you update A0 or A1, you'll notice that A2 automagically updates too.

JavaScript doesn’t usually work like this. If we were to write something comparable in JavaScript:

js
let A0 = 1
let A1 = 2
let A2 = A0 + A1

console.log(A2) // 3

A0 = 2
console.log(A2) // Still 3

When we mutate A0, A2 does not change automatically.

So how would we do this in JavaScript? First, in order to re-run the code that updates A2, let's wrap it in a function:

js
let A2

function update() {
  A2 = A0 + A1
}

Then, we need to define a few terms:

  • The update() function produces a side effect, or effect for short, because it modifies the state of the program.

  • A0 and A1 are considered dependencies of the effect, as their values are used to perform the effect. The effect is said to be a subscriber to its dependencies.

What we need is a magic function that can invoke update() (the effect) whenever A0 or A1 (the dependencies) change:

js
whenDepsChange(update)

This whenDepsChange() function has the following tasks:

  1. Track when a variable is read. E.g. when evaluating the expression A0 + A1, both A0 and A1 are read.

  2. If a variable is read when there is a currently running effect, make that effect a subscriber to that variable. E.g. because A0 and A1 are read when update() is being executed, update() becomes a subscriber to both A0 and A1 after the first call.

  3. Detect when a variable is mutated. E.g. when A0 is assigned a new value, notify all its subscriber effects to re-run.

How Reactivity Works in Vue {#how-reactivity-works-in-vue}

We can't really track the reading and writing of local variables like in the example. There's just no mechanism for doing that in vanilla JavaScript. What we can do though, is intercept the reading and writing of object properties.

There are two ways of intercepting property access in JavaScript: getter / setters and Proxies. Vue 2 used getter / setters exclusively due to browser support limitations. In Vue 3, Proxies are used for reactive objects and getter / setters are used for refs. Here's some pseudo-code that illustrates how they work:

js{4,9,17,22}
function reactive(obj) {
  return new Proxy(obj, {
    get(target, key) {
      track(target, key)
      return target[key]
    },
    set(target, key, value) {
      target[key] = value
      trigger(target, key)
    }
  })
}

function ref(value) {
  const refObject = {
    get value() {
      track(refObject, 'value')
      return value
    },
    set value(newValue) {
      value = newValue
      trigger(refObject, 'value')
    }
  }
  return refObject
}

:::tip
Code snippets here and below are meant to explain the core concepts in the simplest form possible, so many details are omitted, and edge cases ignored.
:::

This explains a few limitations of reactive objects that we have discussed in the fundamentals section:

  • When you assign or destructure a reactive object's property to a local variable, accessing or assigning to that variable is non-reactive because it no longer triggers the get / set proxy traps on the source object. Note this "disconnect" only affects the variable binding - if the variable points to a non-primitive value such as an object, mutating the object would still be reactive.

  • The returned proxy from reactive(), although behaving just like the original, has a different identity if we compare it to the original using the === operator.

Inside track(), we check whether there is a currently running effect. If there is one, we lookup the subscriber effects (stored in a Set) for the property being tracked, and add the effect to the Set:

js
// This will be set right before an effect is about
// to be run. We'll deal with this later.
let activeEffect

function track(target, key) {
  if (activeEffect) {
    const effects = getSubscribersForProperty(target, key)
    effects.add(activeEffect)
  }
}

Effect subscriptions are stored in a global WeakMap<target, Map<key, Set<effect>>> data structure. If no subscribing effects Set was found for a property (tracked for the first time), it will be created. This is what the getSubscribersForProperty() function does, in short. For simplicity, we will skip its details.

Inside trigger(), we again lookup the subscriber effects for the property. But this time we invoke them instead:

js
function trigger(target, key) {
  const effects = getSubscribersForProperty(target, key)
  effects.forEach((effect) => effect())
}

Now let's circle back to the whenDepsChange() function:

js
function whenDepsChange(update) {
  const effect = () => {
    activeEffect = effect
    update()
    activeEffect = null
  }
  effect()
}

It wraps the raw update function in an effect that sets itself as the current active effect before running the actual update. This enables track() calls during the update to locate the current active effect.

At this point, we have created an effect that automatically tracks its dependencies, and re-runs whenever a dependency changes. We call this a Reactive Effect.

Vue provides an API that allows you to create reactive effects: watchEffect(). In fact, you may have noticed that it works pretty similarly to the magical whenDepsChange() in the example. We can now rework the original example using actual Vue APIs:

js
import { ref, watchEffect } from 'vue'

const A0 = ref(0)
const A1 = ref(1)
const A2 = ref()

watchEffect(() => {
  // tracks A0 and A1
  A2.value = A0.value + A1.value
})

// triggers the effect
A0.value = 2

Using a reactive effect to mutate a ref isn't the most interesting use case - in fact, using a computed property makes it more declarative:

js
import { ref, computed } from 'vue'

const A0 = ref(0)
const A1 = ref(1)
const A2 = computed(() => A0.value + A1.value)

A0.value = 2

Internally, computed manages its invalidation and re-computation using a reactive effect.

So what's an example of a common and useful reactive effect? Well, updating the DOM! We can implement simple "reactive rendering" like this:

js
import { ref, watchEffect } from 'vue'

const count = ref(0)

watchEffect(() => {
  document.body.innerHTML = `Count is: ${count.value}`
})

// updates the DOM
count.value++

In fact, this is pretty close to how a Vue component keeps the state and the DOM in sync - each component instance creates a reactive effect to render and update the DOM. Of course, Vue components use much more efficient ways to update the DOM than innerHTML. This is discussed in Rendering Mechanism.

The ref(), computed() and watchEffect() APIs are all part of the Composition API. If you have only been using Options API with Vue so far, you'll notice that Composition API is closer to how Vue's reactivity system works under the hood. In fact, in Vue 3 the Options API is implemented on top of the Composition API. All property access on the component instance (this) triggers getter / setters for reactivity tracking, and options like watch and computed invoke their Composition API equivalents internally.

Runtime vs. Compile-time Reactivity {#runtime-vs-compile-time-reactivity}

Vue's reactivity system is primarily runtime-based: the tracking and triggering are all performed while the code is running directly in the browser. The pros of runtime reactivity are that it can work without a build step, and there are fewer edge cases. On the other hand, this makes it constrained by the syntax limitations of JavaScript, leading to the need of value containers like Vue refs.

Some frameworks, such as Svelte, choose to overcome such limitations by implementing reactivity during compilation. It analyzes and transforms the code in order to simulate reactivity. The compilation step allows the framework to alter the semantics of JavaScript itself - for example, implicitly injecting code that performs dependency analysis and effect triggering around access to locally defined variables. The downside is that such transforms require a build step, and altering JavaScript semantics is essentially creating a language that looks like JavaScript but compiles into something else.

The Vue team did explore this direction via an experimental feature called Reactivity Transform, but in the end we have decided that it would not be a good fit for the project due to the reasoning here.

Reactivity Debugging {#reactivity-debugging}

It's great that Vue's reactivity system automatically tracks dependencies, but in some cases we may want to figure out exactly what is being tracked, or what is causing a component to re-render.

Component Debugging Hooks {#component-debugging-hooks}

We can debug what dependencies are used during a component's render and which dependency is triggering an update using the renderTracked``onRenderTracked and renderTriggered``onRenderTriggered lifecycle hooks. Both hooks will receive a debugger event which contains information on the dependency in question. It is recommended to place a debugger statement in the callbacks to interactively inspect the dependency:

vue
<script setup>
import { onRenderTracked, onRenderTriggered } from 'vue'

onRenderTracked((event) => {
  debugger
})

onRenderTriggered((event) => {
  debugger
})
</script>
js
export default {
  renderTracked(event) {
    debugger
  },
  renderTriggered(event) {
    debugger
  }
}

:::tip
Component debug hooks only work in development mode.
:::

The debug event objects have the following type:

ts
type DebuggerEvent = {
  effect: ReactiveEffect
  target: object
  type:
    | TrackOpTypes /* 'get' | 'has' | 'iterate' */
    | TriggerOpTypes /* 'set' | 'add' | 'delete' | 'clear' */
  key: any
  newValue?: any
  oldValue?: any
  oldTarget?: Map<any, any> | Set<any>
}

Computed Debugging {#computed-debugging}

We can debug computed properties by passing computed() a second options object with onTrack and onTrigger callbacks:

  • onTrack will be called when a reactive property or ref is tracked as a dependency.
  • onTrigger will be called when the watcher callback is triggered by the mutation of a dependency.

Both callbacks will receive debugger events in the same format as component debug hooks:

js
const plusOne = computed(() => count.value + 1, {
  onTrack(e) {
    // triggered when count.value is tracked as a dependency
    debugger
  },
  onTrigger(e) {
    // triggered when count.value is mutated
    debugger
  }
})

// access plusOne, should trigger onTrack
console.log(plusOne.value)

// mutate count.value, should trigger onTrigger
count.value++

:::tip
onTrack and onTrigger computed options only work in development mode.
:::

Computed debugging options are only available via the Composition API computed() function.

Watcher Debugging {#watcher-debugging}

Similar to computed(), watchers also support the onTrack and onTrigger options:

js
watch(source, callback, {
  onTrack(e) {
    debugger
  },
  onTrigger(e) {
    debugger
  }
})

watchEffect(callback, {
  onTrack(e) {
    debugger
  },
  onTrigger(e) {
    debugger
  }
})

Watchers declared with the object syntax also support the onTrack and onTrigger options:

js
export default {
  watch: {
    source: {
      handler() {
        // ...
      },
      onTrack(e) {
        debugger
      },
      onTrigger(e) {
        debugger
      }
    }
  }
}

:::tip
onTrack and onTrigger watcher options only work in development mode.
:::

Integration with External State Systems {#integration-with-external-state-systems}

Vue's reactivity system works by deeply converting plain JavaScript objects into reactive proxies. The deep conversion can be unnecessary or sometimes unwanted when integrating with external state management systems (e.g. if an external solution also uses Proxies).

The general idea of integrating Vue's reactivity system with an external state management solution is to hold the external state in a shallowRef. A shallow ref is only reactive when its .value property is accessed - the inner value is left intact. When the external state changes, replace the ref value to trigger updates.

Immutable Data {#immutable-data}

If you are implementing an undo / redo feature, you likely want to take a snapshot of the application's state on every user edit. However, Vue's mutable reactivity system isn't best suited for this if the state tree is large, because serializing the entire state object on every update can be expensive in terms of both CPU and memory costs.

Immutable data structures solve this by never mutating the state objects - instead, it creates new objects that share the same, unchanged parts with old ones. There are different ways of using immutable data in JavaScript, but we recommend using Immer with Vue because it allows you to use immutable data while keeping the more ergonomic, mutable syntax.

We can integrate Immer with Vue via a simple composable:

js
import { produce } from 'immer'
import { shallowRef } from 'vue'

export function useImmer(baseState) {
  const state = shallowRef(baseState)
  const update = (updater) => {
    state.value = produce(state.value, updater)
  }

  return [state, update]
}

Try it in the Playground

State Machines {#state-machines}

State Machine is a model for describing all the possible states an application can be in, and all the possible ways it can transition from one state to another. While it may be overkill for simple components, it can help make complex state flows more robust and manageable.

One of the most popular state machine implementations in JavaScript is XState. Here's a composable that integrates with it:

js
import { createMachine, interpret } from 'xstate'
import { shallowRef } from 'vue'

export function useMachine(options) {
  const machine = createMachine(options)
  const state = shallowRef(machine.initialState)
  const service = interpret(machine)
    .onTransition((newState) => (state.value = newState))
    .start()
  const send = (event) => service.send(event)

  return [state, send]
}

Try it in the Playground

RxJS {#rxjs}

RxJS is a library for working with asynchronous event streams. The VueUse library provides the @vueuse/rxjs add-on for connecting RxJS streams with Vue's reactivity system.

Connection to Signals {#connection-to-signals}

Quite a few other frameworks have introduced reactivity primitives similar to refs from Vue's Composition API, under the term "signals":

Fundamentally, signals are the same kind of reactivity primitive as Vue refs. It's a value container that provides dependency tracking on access, and side-effect triggering on mutation. This reactivity-primitive-based paradigm isn't a particularly new concept in the frontend world: it dates back to implementations like Knockout observables and Meteor Tracker from more than a decade ago. Vue Options API and the React state management library MobX are also based on the same principles, but hide the primitives behind object properties.

Although not a necessary trait for something to qualify as signals, today the concept is often discussed alongside the rendering model where updates are performed through fine-grained subscriptions. Due to the use of Virtual DOM, Vue currently relies on compilers to achieve similar optimizations. However, we are also exploring a new Solid-inspired compilation strategy, called Vapor Mode, that does not rely on Virtual DOM and takes more advantage of Vue's built-in reactivity system.

API Design Trade-Offs {#api-design-trade-offs}

The design of Preact and Qwik's signals are very similar to Vue's shallowRef: all three provide a mutable interface via the .value property. We will focus the discussion on Solid and Angular signals.

Solid Signals {#solid-signals}

Solid's createSignal() API design emphasizes read / write segregation. Signals are exposed as a read-only getter and a separate setter:

js
const [count, setCount] = createSignal(0)

count() // access the value
setCount(1) // update the value

Notice how the count signal can be passed down without the setter. This ensures that the state can never be mutated unless the setter is also explicitly exposed. Whether this safety guarantee justifies the more verbose syntax could be subject to the requirement of the project and personal taste - but in case you prefer this API style, you can easily replicate it in Vue:

js
import { shallowRef, triggerRef } from 'vue'

export function createSignal(value, options) {
  const r = shallowRef(value)
  const get = () => r.value
  const set = (v) => {
    r.value = typeof v === 'function' ? v(r.value) : v
    if (options?.equals === false) triggerRef(r)
  }
  return [get, set]
}

Try it in the Playground

Angular Signals {#angular-signals}

Angular is undergoing some fundamental changes by foregoing dirty-checking and introducing its own implementation of a reactivity primitive. The Angular Signal API looks like this:

js
const count = signal(0)

count() // access the value
count.set(1) // set new value
count.update((v) => v + 1) // update based on previous value

Again, we can easily replicate the API in Vue:

js
import { shallowRef } from 'vue'

export function signal(initialValue) {
  const r = shallowRef(initialValue)
  const s = () => r.value
  s.set = (value) => {
    r.value = value
  }
  s.update = (updater) => {
    r.value = updater(r.value)
  }
  return s
}

Try it in the Playground

Compared to Vue refs, Solid and Angular's getter-based API style provide some interesting trade-offs when used in Vue components:

  • () is slightly less verbose than .value, but updating the value is more verbose.
  • There is no ref-unwrapping: accessing values always require (). This makes value access consistent everywhere. This also means you can pass raw signals down as component props.

Whether these API styles suit you is to some extent subjective. Our goal here is to demonstrate the underlying similarity and trade-offs between these different API designs. We also want to show that Vue is flexible: you are not really locked into the existing APIs. Should it be necessary, you can create your own reactivity primitive API to suit more specific needs.


File: src/guide/extras/reactivity-transform.md

Reactivity Transform {#reactivity-transform}

:::danger Removed Experimental Feature
Reactivity Transform was an experimental feature, and has been removed in the latest 3.4 release. Please read about the reasoning here.

If you still intend to use it, it is now available via the Vue Macros plugin.
:::

:::tip Composition-API-specific
Reactivity Transform is a Composition-API-specific feature and requires a build step.
:::

Refs vs. Reactive Variables {#refs-vs-reactive-variables}

Ever since the introduction of the Composition API, one of the primary unresolved questions is the use of refs vs. reactive objects. It's easy to lose reactivity when destructuring reactive objects, while it can be cumbersome to use .value everywhere when using refs. Also, .value is easy to miss if not using a type system.

Vue Reactivity Transform is a compile-time transform that allows us to write code like this:

vue
<script setup>
let count = $ref(0)

console.log(count)

function increment() {
  count++
}
</script>

<template>
  {{ count }}
</template>

The $ref() method here is a compile-time macro: it is not an actual method that will be called at runtime. Instead, the Vue compiler uses it as a hint to treat the resulting count variable as a reactive variable.

Reactive variables can be accessed and re-assigned just like normal variables, but these operations are compiled into refs with .value. For example, the <script> part of the above component is compiled into:

js{5,8}
import { ref } from 'vue'

let count = ref(0)

console.log(count.value)

function increment() {
  count.value++
}

Every reactivity API that returns refs will have a $-prefixed macro equivalent. These APIs include:

These macros are globally available and do not need to be imported when Reactivity Transform is enabled, but you can optionally import them from vue/macros if you want to be more explicit:

js
import { $ref } from 'vue/macros'

let count = $ref(0)

Destructuring with `$()` {#destructuring-with}

It is common for a composition function to return an object of refs, and use destructuring to retrieve these refs. For this purpose, reactivity transform provides the $() macro:

js
import { useMouse } from '@vueuse/core'

const { x, y } = $(useMouse())

console.log(x, y)

Compiled output:

js
import { toRef } from 'vue'
import { useMouse } from '@vueuse/core'

const __temp = useMouse(),
  x = toRef(__temp, 'x'),
  y = toRef(__temp, 'y')

console.log(x.value, y.value)

Note that if x is already a ref, toRef(__temp, 'x') will simply return it as-is and no additional ref will be created. If a destructured value is not a ref (e.g. a function), it will still work - the value will be wrapped in a ref so the rest of the code works as expected.

$() destructure works on both reactive objects and plain objects containing refs.

Convert Existing Refs to Reactive Variables with `$()` {#convert-existing-refs-to-reactive-variables-with}

In some cases we may have wrapped functions that also return refs. However, the Vue compiler won't be able to know ahead of time that a function is going to return a ref. In such cases, the $() macro can also be used to convert any existing refs into reactive variables:

js
function myCreateRef() {
  return ref(0)
}

let count = $(myCreateRef())

Reactive Props Destructure {#reactive-props-destructure}

There are two pain points with the current defineProps() usage in <script setup>:

  1. Similar to .value, you need to always access props as props.x in order to retain reactivity. This means you cannot destructure defineProps because the resulting destructured variables are not reactive and will not update.

  2. When using the type-only props declaration, there is no easy way to declare default values for the props. We introduced the withDefaults() API for this exact purpose, but it's still clunky to use.

We can address these issues by applying a compile-time transform when defineProps is used with destructuring, similar to what we saw earlier with $():

html
<script setup lang="ts">
  interface Props {
    msg: string
    count?: number
    foo?: string
  }

  const {
    msg,
    // default value just works
    count = 1,
    // local aliasing also just works
    // here we are aliasing `props.foo` to `bar`
    foo: bar
  } = defineProps()

  watchEffect(() => {
    // will log whenever the props change
    console.log(msg, count, bar)
  })
</script>

The above will be compiled into the following runtime declaration equivalent:

js
export default {
  props: {
    msg: { type: String, required: true },
    count: { type: Number, default: 1 },
    foo: String
  },
  setup(props) {
    watchEffect(() => {
      console.log(props.msg, props.count, props.foo)
    })
  }
}

Retaining Reactivity Across Function Boundaries {#retaining-reactivity-across-function-boundaries}

While reactive variables relieve us from having to use .value everywhere, it creates an issue of "reactivity loss" when we pass reactive variables across function boundaries. This can happen in two cases:

Passing into function as argument {#passing-into-function-as-argument}

Given a function that expects a ref as an argument, e.g.:

ts
function trackChange(x: Ref<number>) {
  watch(x, (x) => {
    console.log('x changed!')
  })
}

let count = $ref(0)
trackChange(count) // doesn't work!

The above case will not work as expected because it compiles to:

ts
let count = ref(0)
trackChange(count.value)

Here count.value is passed as a number, whereas trackChange expects an actual ref. This can be fixed by wrapping count with $$() before passing it:

diff
let count = $ref(0)
- trackChange(count)
+ trackChange($$(count))

The above compiles to:

js
import { ref } from 'vue'

let count = ref(0)
trackChange(count)

As we can see, $$() is a macro that serves as an escape hint: reactive variables inside $$() will not get .value appended.

Returning inside function scope {#returning-inside-function-scope}

Reactivity can also be lost if reactive variables are used directly in a returned expression:

ts
function useMouse() {
  let x = $ref(0)
  let y = $ref(0)

  // listen to mousemove...

  // doesn't work!
  return {
    x,
    y
  }
}

The above return statement compiles to:

ts
return {
  x: x.value,
  y: y.value
}

In order to retain reactivity, we should be returning the actual refs, not the current value at return time.

Again, we can use $$() to fix this. In this case, $$() can be used directly on the returned object - any reference to reactive variables inside the $$() call will retain the reference to their underlying refs:

ts
function useMouse() {
  let x = $ref(0)
  let y = $ref(0)

  // listen to mousemove...

  // fixed
  return $$({
    x,
    y
  })
}

Using `$$()` on destructured props {#using-on-destructured-props}

$$() works on destructured props since they are reactive variables as well. The compiler will convert it with toRef for efficiency:

ts
const { count } = defineProps<{ count: number }>()

passAsRef($$(count))

compiles to:

js
setup(props) {
  const __props_count = toRef(props, 'count')
  passAsRef(__props_count)
}

TypeScript Integration {#typescript-integration}

Vue provides typings for these macros (available globally) and all types will work as expected. There are no incompatibilities with standard TypeScript semantics, so the syntax will work with all existing tooling.

This also means the macros can work in any files where valid JS / TS are allowed - not just inside Vue SFCs.

Since the macros are available globally, their types need to be explicitly referenced (e.g. in a env.d.ts file):

ts
/// <reference types="vue/macros-global" />

When explicitly importing the macros from vue/macros, the type will work without declaring the globals.

Explicit Opt-in {#explicit-opt-in}

:::danger No longer supported in core
The following only applies up to Vue version 3.3 and below. Support has been removed in Vue core 3.4 and above, and @vitejs/plugin-vue 5.0 and above. If you intend to continue using the transform, please migrate to Vue Macros instead.
:::

Vite {#vite}

  • Requires @vitejs/plugin-vue@>=2.0.0
  • Applies to SFCs and js(x)/ts(x) files. A fast usage check is performed on files before applying the transform so there should be no performance cost for files not using the macros.
  • Note reactivityTransform is now a plugin root-level option instead of nested as script.refSugar, since it affects not just SFCs.
js [vite.config.js]
export default {
  plugins: [
    vue({
      reactivityTransform: true
    })
  ]
}

`vue-cli` {#vue-cli}

  • Currently only affects SFCs
  • Requires vue-loader@>=17.0.0
js [vue.config.js]
module.exports = {
  chainWebpack: (config) => {
    config.module
      .rule('vue')
      .use('vue-loader')
      .tap((options) => {
        return {
          ...options,
          reactivityTransform: true
        }
      })
  }
}

Plain `webpack` + `vue-loader` {#plain-webpack-vue-loader}

  • Currently only affects SFCs
  • Requires vue-loader@>=17.0.0
js [webpack.config.js]
module.exports = {
  module: {
    rules: [
      {
        test: /\.vue$/,
        loader: 'vue-loader',
        options: {
          reactivityTransform: true
        }
      }
    ]
  }
}

File: src/guide/extras/render-function.md


outline: deep

Render Functions & JSX {#render-functions-jsx}

Vue recommends using templates to build applications in the vast majority of cases. However, there are situations where we need the full programmatic power of JavaScript. That's where we can use the render function.

If you are new to the concept of virtual DOM and render functions, make sure to read the Rendering Mechanism chapter first.

Basic Usage {#basic-usage}

Creating Vnodes {#creating-vnodes}

Vue provides an h() function for creating vnodes:

js
import { h } from 'vue'

const vnode = h(
  'div', // type
  { id: 'foo', class: 'bar' }, // props
  [
    /* children */
  ]
)

h() is short for hyperscript - which means "JavaScript that produces HTML (hypertext markup language)". This name is inherited from conventions shared by many virtual DOM implementations. A more descriptive name could be createVNode(), but a shorter name helps when you have to call this function many times in a render function.

The h() function is designed to be very flexible:

js
// all arguments except the type are optional
h('div')
h('div', { id: 'foo' })

// both attributes and properties can be used in props
// Vue automatically picks the right way to assign it
h('div', { class: 'bar', innerHTML: 'hello' })

// props modifiers such as `.prop` and `.attr` can be added
// with `.` and `^` prefixes respectively
h('div', { '.name': 'some-name', '^width': '100' })

// class and style have the same object / array
// value support that they have in templates
h('div', { class: [foo, { bar }], style: { color: 'red' } })

// event listeners should be passed as onXxx
h('div', { onClick: () => {} })

// children can be a string
h('div', { id: 'foo' }, 'hello')

// props can be omitted when there are no props
h('div', 'hello')
h('div', [h('span', 'hello')])

// children array can contain mixed vnodes and strings
h('div', ['hello', h('span', 'hello')])

The resulting vnode has the following shape:

js
const vnode = h('div', { id: 'foo' }, [])

vnode.type // 'div'
vnode.props // { id: 'foo' }
vnode.children // []
vnode.key // null

:::warning Note
The full VNode interface contains many other internal properties, but it is strongly recommended to avoid relying on any properties other than the ones listed here. This avoids unintended breakage in case the internal properties are changed.
:::

Declaring Render Functions {#declaring-render-functions}

When using templates with Composition API, the return value of the setup() hook is used to expose data to the template. When using render functions, however, we can directly return the render function instead:

js
import { ref, h } from 'vue'

export default {
  props: {
    /* ... */
  },
  setup(props) {
    const count = ref(1)

    // return the render function
    return () => h('div', props.msg + count.value)
  }
}

The render function is declared inside setup() so it naturally has access to the props and any reactive state declared in the same scope.

In addition to returning a single vnode, you can also return strings or arrays:

js
export default {
  setup() {
    return () => 'hello world!'
  }
}
js
import { h } from 'vue'

export default {
  setup() {
    // use an array to return multiple root nodes
    return () => [
      h('div'),
      h('div'),
      h('div')
    ]
  }
}

:::tip
Make sure to return a function instead of directly returning values! The setup() function is called only once per component, while the returned render function will be called multiple times.
:::

We can declare render functions using the render option:

js
import { h } from 'vue'

export default {
  data() {
    return {
      msg: 'hello'
    }
  },
  render() {
    return h('div', this.msg)
  }
}

The render() function has access to the component instance via this.

In addition to returning a single vnode, you can also return strings or arrays:

js
export default {
  render() {
    return 'hello world!'
  }
}
js
import { h } from 'vue'

export default {
  render() {
    // use an array to return multiple root nodes
    return [
      h('div'),
      h('div'),
      h('div')
    ]
  }
}

If a render function component doesn't need any instance state, they can also be declared directly as a function for brevity:

js
function Hello() {
  return 'hello world!'
}

That's right, this is a valid Vue component! See Functional Components for more details on this syntax.

Vnodes Must Be Unique {#vnodes-must-be-unique}

All vnodes in the component tree must be unique. That means the following render function is invalid:

js
function render() {
  const p = h('p', 'hi')
  return h('div', [
    // Yikes - duplicate vnodes!
    p,
    p
  ])
}

If you really want to duplicate the same element/component many times, you can do so with a factory function. For example, the following render function is a perfectly valid way of rendering 20 identical paragraphs:

js
function render() {
  return h(
    'div',
    Array.from({ length: 20 }).map(() => {
      return h('p', 'hi')
    })
  )
}

Using Vnodes in `