Miss any of our Open RFC calls?Watch the recordings here! »


3.3.3 • Public • Published

The next generation of Markdown editor, built for the future

npm bundle size

中文  |  Demo

💡 Introduction

Vditor is a browser-side Markdown editor, implemented using TypeScript. Support native JavaScript, Vue, React and Angular, provide desktop.

Welcome to Vditor Official Site to learn more.

📽️ Background

In the initial stage of developing Sym, we directly used WYSIWYG rich text editor. At that time, HTML-based editors were very popular, and it was very convenient to quote them in the project, which also conformed to the usage habits of users at that time.

Later, the rise of Markdown gradually changed everyone's typography. In addition, several of our other projects are for programmer users, so it is also a general trend to migrate to md. We chose CodeMirror, which is an excellent editor, it provides a rich programming interface for developers, and is also compatible with various browsers. it is good.

Later, as the business needs of our projects have precipitated, using CodeMirror sometimes feels more "cumbersome." For example, to implement @automatically complete the user name list, insert Emoji, upload files, etc., it requires more in-depth secondary development, and these business requirements are precisely common and necessary in many project scenarios.

Finally, we decided to start implementing the editor ourselves in Sym. With the iterations of several versions, Sym's editor has matured. In the community HacPai that we operate, people have asked us if we can separate the editor for everyone to use. At the same time, our front-end main programmer V also felt a little bit overwhelmed with maintaining the editors scattered in various projects, plus a good impression of TypeScript, so I decided Use ts to implement a new browser-side md editor.

So, Vditor was born.

✨ Features

  • Support three editing modes: WYSIWYG(wysiwyg), Instant Rendering(ir) and Split View(sv)
  • Support outline, mathematical formulas, mind maps, charts, flowcharts, Gantt charts, timing charts, staffs, multimedia, voice reading, heading anchors, code highlighting and copying, graphviz rendering
  • Built-in security filtering, export, image lazy loading, task list, at, multi-platform preview, multi-theme switching, copy to WeChat function
  • Implementation of CommonMark and GFM specifications, formatting and syntax tree viewing of Markdown, and support for 10+ configurations
  • The toolbar contains 36+ items of operations. In addition to support for expansion, the shortcut keys, tip, tip positions, icons, click events, class names, and sub-toolbars can be customized
  • Emoji auto-complete, set common emoticons, support emoticon customization
  • You can use drag and drop, paste and paste to upload, display real-time upload progress, support CORS cross-domain upload
  • Save content in real time to prevent accidental loss
  • Recording support, users can directly publish voice
  • The markup HTML is automatically converted to Markdown, if the paste contains images of external links, it can be uploaded to the server through the specified interface
  • Support main window size drag and drop, character counting
  • Multi-theme support, built-in black, white and green three sets of themes
  • Multi-language support, built-in Chinese, English, Korean text localization
  • Support mainstream browsers, mobile friendly



🔮 Editing Modes


WYSIWYG mode is more friendly to users who are not familiar with Markdown, and you can use it seamlessly if you are familiar with Markdown.


Instant Rendering

Instant Rendering mode should not be unfamiliar to users who are familiar with Typora. In theory, this is the most elegant Markdown editing method.


Split View

The traditional Split View mode is suitable for Markdown editing on a large screen.


🗃 Showcases

  • 🎶 Sym A modern community (forum/BBS/SNS/blog) platform implemented in Java
  • 🎸 Solo & 🎷 Pipe B3log distributed community blog end node, welcome to join the next generation community network
  • 📕 LianDi Note A desktop note application that supports Windows, Mac and Linux
  • 🌟 Starfire A distributed content-sharing and discussion community, the spark can catch fire
  • 📝 Arya Based on Vue, Vditor, built online Markdown editor

🛠️ User Guide


  • Install dependencies
npm install vditor --save
  • Introduce and initialize objects in the code, you can refer to index.js
import Vditor from 'vditor'
import "~vditor/src/assets/scss/index" // Or use dark
const vditor = new Vditor(id, {options...})

HTML script

  • Insert CSS and js in HTML, you can refer to demo
<!-- ⚠️Please specify the version number in the production environment, such as https://cdn.jsdelivr.net/npm/vditor@x.x.x/dist... -->
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/vditor/dist/index.css" />
<script src="https://cdn.jsdelivr.net/npm/vditor/dist/index.min.js" defer></script>

Demo code


  • Support two sets of black and white themes: classic/dark
  • Use the scss/css developed by yourself to fully customize the style after referring to the existing style
  • Theme colors can be customized by modifying variables in index.scss
  • Adding class="vditor-reset" (classic theme) or class="vditor-reset vditor-reset--dark" (black theme) attribute on the content display element can display the content more friendly



Can be filled with element id or element itselfHTMLElement

⚠️: When filling in the element's HTMLElement, you need to set options.cache.id or set options.cache.enable to false


Explanation Default
after Callback method after editor asynchronous rendering is completed -
height Total editor height 'auto'
minHeight Editing area minimum height -
width Total editor width, supports % 'auto'
placeholder Tips when the input area is empty ''
lang i18n: en_US, ko_KR, zh_CN 'zh_CN'
input Trigger after input (value: string, previewElement?: HTMLElement): void -
focus Trigger after focusing (value: string): void -
blur Trigger after out of focus (value: string): void -
esc Trigger after pressing esc (value: string): void -
ctrlEnter Trigger after pressing ⌘/ctrl+enter (value: string): void -
select Triggered after selecting text in the editor (value: string): void -
tab tab key operation string, support \ t and any string -
typewriterMode Whether to enable typewriter mode false
cdn Configure self-built CDN address https://cdn.jsdelivr.net/npm/vditor@${VDITOR_VERSION}
mode Editing mode: sv, ir, wysiwyg 'wysiwyg'
debugger Whether to display the log false
value Editor initialization value ''
theme Theme: classic, dark 'classic'
outline show outline false


  • Toolbar, you can use name for shorthand: toolbar: ['emoji', 'br', 'bold', '|', 'line']. See default src/ts/util/Options.ts
  • name can be enumerated as: emoji , headings , bold , italic , strike , | , line , quote , list , ordered-list , check ,outdent ,indent , code , inline-code, insert-after, insert-before, code-theme, content-theme, export, undo , redo , upload , link , table , record , edit-mode , both , preview , format , fullscreen , outline , devtools , info , help , br
  • When name is not in the enumeration, you can add a custom button in the following format:
 hotkey: '⌘-⇧-f',  
 name: 'format',  
 tipPosition: 'ne',  
 tip: 'format',  
 className: '',
 icon: '<svg version="1.1" xmlns="http://www.w3.org/2000/svg" width="768" height="768" viewBox="0 0 768 768"><path d="M342 426v-84h426v84h-426zM342 256v-86h426v86h-426zM0 0h768v86h-768v-86zM342 598v-86h426v86h-426zM0 214l170 170-170 170v-340zM0 768v-86h768v86h-768z"></path></svg>',  
 click: () => {  
   alert('custom toolbar')  
Explanation Default
name Unique label -
icon svg icon -
tip Prompt -
tipPosition Prompt location: ne, nw -
hotkey Shortcut keys, support ⌘/ctrl-key or ⌘/ctrl-⇧/shift-key format configuration, do not support wysiwyg mode -
suffix Insert the suffix in the editor -
prefix Insert the prefix in the editor -
click Custom event triggered when button is clicked (): void -
className Style name ''
toolbar?: Array<options.toolbar> sub menu -


Explanation Default
hide Whether to hide the toolbar false
pin Whether to pin the toolbar false


Explanation Default
enable Whether to use counter false
max max counter -
type counter type: md, text 'md'


Explanation Default
enable Whether to use localStorage for caching true
id Cache key, the first parameter is an element and when caching is enabled required -
after cache callback (markdown: string): void -


Explanation Default
delay Preview debounce millisecond interval 1000
maxWidth Preview area maximum width 800
mode Display mode: both, editor 'both'
url md parsing request -
parse Preview callback (element: HTMLElement): void -
transform Callback before rendering (html: string): string -


Explanation Default
enable Whether to enable code syntax highlighting true
style For optional values, see Chroma github
lineNumber Whether to enable line number false


Explanation Default
autoSpace Autospace false
fixTermTypo Automatically correct terminology false
chinesePunct Automatic punctuation correction false
toc Insert Table of Contents false
footnotes Footnotes true
codeBlockPreview Whether to render code blocks in wysiwyg and ir modes true
theme Content Theme 'light'
setext Whether to parse the setext header fa;se
paragraphBeginningSpace Two spaces before the paragraph false
sanitize Use XSS true
listStyle add data-style attribute false
linkBase link prefix ''


Explanation Default
inlineDigit Whether numbers are allowed after the inline math formula starting with $ false
macros Macro definition passed in when rendering with MathJax {}
engine Math formula rendering engine: KaTeX, MathJax 'KaTeX'


Explanation Default
delay Tip debounce millisecond interval 200
emoji The default emoji can be selected from lute/emoji_map, or can be customized { '+1': '👍', '-1': '👎', 'heart': '❤️', 'cold_sweat': '😰' }
emojiTail Common emoji -
emojiPath Emoji path https://cdn.jsdelivr.net/npm/vditor@${VDITOR_VERSION}/dist/images/emoji
at @user callback (value: string): Array<any>, Need to return array synchronously [{value: '', html: ''}] -


  • The data structure of the file upload is as follows. When the data structure returned by the backend is inconsistent, you can use format for conversion.
// POST data  
xhr.send(formData);  // formData = FormData.append("file[]", File)  
// return data  
 "msg": "",  
 "code": 0,  
 "data": {  
 "errFiles": ['filename', 'filename2'],  
 "succMap": {  
   "filename3": "filepath3",  
   "filename3": "filepath3"  
  • In order to prevent the off-site pictures from being invalid, linkToImgUrl can transfer the off-site picture addresses in the clipboard to the server for saving and processing. The data structure is as follows:
// POST data  
xhr.send(JSON.stringify({url: src})); // src is the address of the image outside the station
// return data  
 msg: '',  
 code: 0,  
 data : {  
   originalURL: '',  
   url: ''  
Explanation Default
url Upload url ''
max The largest upload file Byte 10 * 1024 * 1024
linkToImgUrl When the clipboard contains the image address, use this url to re-upload ''
success Upload success callback (editor: HTMLPreElement, msg: string): void -
error Upload failure callback (msg: string): void -
token CORS upload verification, header is X-Upload-Token -
withCredentials Cross-site access control false
headers Request header settings -
filename Sanitizing file names (name: string): string | name => name.replace(/\W/g, '')
accept File upload type, same as input accept -
validate Check, return true if successful, otherwise return error message (files: File[]) => string | boolean -
handler Custom upload, return error message when an error occurs (files: File[]) => string | null -
format Transform the data returned by the server to meet the built-in data structure (files: File[], responseText: string): string -
file Process the uploaded file before returning (files: File[]): File[] -
setHeaders Use the return value to set the header before uploading (): { [key: string]: string } -
extraData Append data to FormData { [key: string]: string Blob }


Explanation Default
enable Whether to support size drag false
position Drag column position:top, bottom 'bottom'
after Callback when dragging ends (height: number): void -


Explanation Default
preview Preview on the element className ''


Explanation Default
deleteLine Delete the cursor line or selected line '⌘-Backspace'
duplicate Copy current line or selected content '⌘-D'


getValue() Get editor content
getHTML() Get preview area content
insertValue(value: string, render = true) Insert content at the focus and markdown rendering by default
focus() Focus on the editor
blur() Make the editor out of focus
disabled() Disable editor
enable() Unedit editor
setSelection(start: number, end: number) Select the string from start to end and does not support wysiwyg mode
getSelection(): string Returns the selected string
setValue(markdown: string, clearStack = false) Set editor content
clearStack() remove undo and redo stack
renderPreview(value?: string) Set preview area content
getCursorPosition():{top: number, left: number} Get focus position
deleteValue() Delete selected content
updateValue(value: string) Update selected content
isUploading() Whether the upload is still in progress
clearCache() clear cache
disabledCache() Disable cache
enableCache() Enable caching
html2md(value: string) HTML to md
tip(text: string, time: number) notification. time is 0 will always display
setPreviewMode(mode: "both" | "editor") Set preview mode
setTheme(theme: "dark" | "classic", contentTheme?: string, codeTheme?: string) Set theme
getCurrentMode(): string Get the editor's current editing mode

static methods

  • When no editing operation is required, just introduce method.min.js and directly call
import VditorPreview from 'vditor/dist/method.min'  
  • When you need to render Markdown on the page, you can directly call the preview method with the following parameters:
previewElementHTMLDivElement,   // Use this element for rendering
markdownstring,  // The original markdown to be rendered
options?: IPreviewOptions {  
 anchor?: number;  // 0: no render, 1: render left, 2: render right
 customEmoji?: { [keystring]string };    // Custom emoji, default is {}
 lang?: (keyof II18nLang);    // Language, default is 'zh_CN'  
 emojiPath?: string;    // Emoji picture path 
 hljs?: IHljs; // Refer to options.preview.hljs 
 speech?: {  // Read the selected content
  enable?: boolean,
 math?: IMath; // Math formula rendering configuration
 transform?(html: string): string// Callback method before rendering
 after?()void// Callback method after rendering
 cdn?: string// Self-built CDN address
 lazyLoadImage?: string// use "https://cdn.jsdelivr.net/npm/vditor/dist/images/img-loading.svg" to lazy load image
 markdown?: options.preview.markdown;
 renderers?: ILuteRender// Custom rendering method https://hacpai.com/article/1588412297062
  • ⚠️method.min.js andindex.min.js cannot be introduced at the same time
mermaidRender(element: HTMLElement, className = ".language-mermaid", cdn = options.cdn) Convert elements in class to className in element to flowchart/sequence diagram/gantt diagram
codeRender(element: HTMLElement, lang: (keyof II18nLang) = "zh_CN") Add a copy button for the code block in element
chartRender(element: (HTMLElement| Document) = document, cdn = options.cdn) Chart rendering
abcRender(element: (HTMLElement| Document) = document, cdn = options.cdn) Stave rendering
outlineRender(contentElement: HTMLElement, targetElement: Element, vditor?: IVditor) Outline rendering
md2html(mdText: string, options?: IPreviewOptions): Promise<string> Markdown text is converted to HTML, this method needs to use asynchronous programming
preview(previewElement: HTMLDivElement, markdown: string, options?: IPreviewOptions) Page Markdown article rendering
highlightRender(hljsOption?: IHljs, element?: HTMLElement | Document, cdn = options.cdn) Highlight the code block in element
mediaRender(element: HTMLElement) Rendering as specific link as video, audio, embedded iframe
mathRender(element: HTMLElement, options?: {cdn?: string, math?: IMath}) Render math formulas
speechRender(element: HTMLElement, lang?: (keyof II18nLang)) Read the selected text
graphvizRender(element: HTMLElement, cdn?: string) Render graphviz
lazyLoadImageRender(element: (HTMLElement | Document) = document) Render lazy load image
setCodeTheme (codeTheme: string, cdn = options.cdn) update code theme
setContentTheme (contentTheme: string, cdn = options.cdn) update content theme
mindmapRender (element: (HTMLElement | Document) = document, cdn = options.cdn) Render mind map

🏗 Developer Guide

Principle related


  1. Install node LTS version
  2. Download latest code and unzip
  3. Run npm install in the root directory
  4. npm run start Start the local server, open http: // localhost: 9000
  5. Modify the code
  6. npm run build package code to dist directory

CDN switch

Due to the on-demand loading mechanism, the default CDN is https://cdn.jsdelivr.net/npm/vditor@version number

If the code is modified or you need to use a self-built CDN, you can follow the steps below:

  • The initial options andIPreviewOptions need to add cdn configuration
  • highlightRender,mathRender, abcRender,chartRender, mermaidRender methods need to add cdn parameter
  • Copy the dist directory in the successful build or jsDelivr to the correct location


Please read CHANGELOG carefully when upgrading the version.

Ⓜ️ Markdown User Guide

🏘️ Community

📄 License

Vditor uses the MIT open source license.

🙏 Acknowledgement

  • Lute: A structured Markdown engine that supports Go and JavaScript
  • highlight.js: JavaScript syntax highlighter
  • mermaid: Generation of diagram and flowchart from text in a similar manner as Markdown
  • incubator-echarts: A powerful, interactive charting and visualization library for browser
  • abcjs: JavaScript library for rendering standard music notation in a browser
  • IntelliJ IDEA: Family of capable and ergonomic development environments


npm i vditor

DownloadsWeekly Downloads






Unpacked Size

12.2 MB

Total Files


Last publish


  • avatar