editorjs-parser is a NPM package for parsing the output object of EditorJs to HTML.
- https://cdn.jsdelivr.net/npm/editorjs-parser@2/dist/index.cjs (Node only)
- https://cdn.jsdelivr.net/npm/editorjs-parser@2/dist/index.global.js (Browser only)
Use the package manager npm to install editorjs-parser.
npm install --save editorjs-parserThe package is written in TypeScript and ships its own type declarations — no
@types install needed. Public types: EditorJSOutput, EditorJSBlock,
ParserConfig, DeepPartial, CustomParsers, CustomEmbeds, ParserFunction.
import edjsParser, { type EditorJSOutput } from "editorjs-parser";
const parser = new edjsParser();
const html: string = parser.parse(output satisfies EditorJSOutput);To use the package in browser, import Browser version through CDN to your HTML file and just call edjsParser class:
const parser = new edjsParser(config, customParsers, embedMarkup);To import the package in Node and Front-end code:
const edjsParser = require("editorjs-parser");
const parser = new edjsParser(config, customParsers, embedMarkup);NOTE: Parameters are optional. If you want to only pass the second parameter, set the first parameter to undefined.
To parse all blocks, pass the exact EditorJs' output object:
const markup = parser.parse(output);To parse one block, pass a complete block:
const markup = parser.parseBlock(block);NOTE: HTML markup in code blocks are already sanitized and ready to be sent to browser. You don't have to do anything.
NOTE: Code blocks are compatible with highlight.js
- Paragraph
- Header
- Table (optional first-row headings via
withHeadings→<thead>/<th>) - Raw
- Delimiter
- Code
- Quote
- List (legacy flat format, and the @editorjs/list v2 format with recursive nesting and checklist style)
- Embed
- Image
- Simple-image
- Checklist
- Warning
- Link tool
- Attaches
- Personality
- Alert
- Block tunes: text variant (
citation/call-out/details), footnotes, and alignment
editor.js saves block tunes in block.tunes alongside block.data. Two
official tunes are rendered:
- Text variant (
@editorjs/text-variant-tune) — the block's HTML is wrapped exactly like the editor renders it:<div class="cdx-text-variant cdx-text-variant--citation">…</div>(same for--call-out,--details). - Footnotes (
@editorjs/footnotes-tune) — the saved footnote texts (matching the<sup data-tune="footnotes">N</sup>markers in the text) are appended as a list after the block:<ol class="cdx-footnotes"><li class="cdx-footnotes__item">…</li></ol>
Third-party alignment tunes (e.g.
editor-js-alignment-tune)
save { alignment: "left|center|right" } under a tune key you choose in your
editor config (commonly anyTuneName), so the parser detects any tune value
shaped { alignment: … } and wraps the block in
<div style="text-align: …">…</div> — the tune defines no output markup of
its own, so this wrapper is our documented convention (like cdx-footnotes).
By default the delimiter block renders the historical <br />. For a semantic
horizontal rule instead:
const parser = new edjsParser({ delimiter: { tag: "hr" } });Blocks added in 1.6.0 render with the official editor.js tools' own CSS class names, so existing editor.js styles apply to parsed output too:
- Warning —
<div class="cdx-warning"><div class="cdx-warning__title">…</div><div class="cdx-warning__message">…</div></div> - Checklist —
<div class="cdx-checklist"><div class="cdx-checklist__item[ cdx-checklist__item--checked]">…</div>…</div> - Link tool —
<a class="link-tool__content link-tool__content--rendered" href="…" target="_blank" rel="nofollow noindex noreferrer">with optionallink-tool__image(background image),link-tool__title,link-tool__descriptionand alink-tool__anchorspan holding the hostname - Attaches —
<div class="cdx-attaches cdx-attaches--with-file">with acdx-attaches__download-buttonlink,cdx-attaches__file-icon(-background/-label),cdx-attaches__titleand a human-readablecdx-attaches__size(KiB/MiB) - Personality —
<div class="cdx-personality">withcdx-personality__photo(background image),cdx-personality__name(linked) andcdx-personality__description - Alert (3.2.0, editorjs-alert) —
<div class="cdx-alert cdx-alert-{type}[ cdx-alert-align-{align}]">with acdx-alert__messagediv; the align class is omitted when the saved data has noalign(tool versions before 1.1)
If you have a custom block like so:
{
type: "customBlock",
data: {
// Your data
}
}You can pass an object of custom parsers or override existing parsers of supported blocks as the second argument, like so:
const customParsers = {
customBlock: function(data, config) {
// parsing functionality
// the config arg is user provided config merged with default config
},
image: function(data, config): {
return `<img src="${data.file.url}" alt="${data.caption}" >`;
}
}
const parser = new edjsParser(undefined, customParsers);NOTE: The config arg is user provided config merged with default configuration.
This is the default configuration. You can override any of these properties by passing a config object.
{
image: {
use: "figure",
// use figure or img tag for images (figcaption will be used for caption of figure)
// if you use figure, caption will be visible
imgClass: "img", // used class for img tags
figureClass: "fig-img", // used class for figure tags
figCapClass: "fig-cap", // used class for figcaption tags
path: "absolute",
// if absolute is passed, the url property which is the absolute path to the image will be used
// otherwise pass a relative path with the filename property in <> like so: '/img/<fileName>'
},
paragraph: {
pClass: "paragraph", // used class for paragraph tags
},
code: {
codeBlockClass: "code-block", // used class for code blocks
},
embed: {
useProvidedLength: false,
// set to true if you want the returned width and height of editorjs to be applied
// NOTE: sometimes source site overrides the lengths so it does not work 100%
},
quote: {
applyAlignment: false,
// if set to true blockquote element will have text-align css property set
},
delimiter: {
tag: "br",
// "hr" renders <hr /> instead of the default <br />
},
};To use the relative path, you should return the filename of the uploaded image from your backend, alongside the url (for more info docs).
Then include the property name of filename in config like so: (for example the property name of the returned filename is imageFileName)
const config = {
image: {
path: "/img/<imageFileName>";
}
};
const parser = new edjsParser(config);NOTE: It is pointless to use both image and simple-image block types in the same editor instance, but this parser supports both of them and you can use any of them that fulfills your needs.
NOTE: Images will have class img.
NOTE: If the image is stretched, the parsed img tag will have img-fullwidth as class.
NOTE: If image is set to have a border, the parsed img tag will have img-border as class.
NOTE: If withBackground is set to true, the parsed img tag will have img-bg as class.
You can style, according to these classes.
If you want the returned width and height of embedded element to be applied, set useProvidedLength option to true in config:
const config = {
embed: {
useProvidedLength: true,
},
};
const parser = new edjsParser(config);If you want to render a custom markup for your embed service, pass it in an object in third argument. For example if you want your own markup to be rendered for Youtube video embed, you got to do this:
const parser = new edjsParser(undefined, undefined, {
youtube: `Your markup in string`,
});You also have access to data object. To use that you should put variable names in placeholders, like so:
const customEmbeds = {
youtube: `<iframe src="<%data.embed%>" width="<%data.width%>"><%data.caption%></iframe>`,
};
const parser = new edjsParser(undefined, undefined, customEmbeds);NOTE: If you want to have useProvidedLength functionality, use <%data.length%> instead of <%data.width%> and <%data.height%> in embed markups.
<%data.length%> returns string like this: width="300" height="500"
If you need the returned alignment of blockquotes to be set, set applyAlignment to true in config:
const config = {
quote: {
applyAlignment: true;
}
};
const parser = new edjsParser(config);npm install
npm test # vitest suite — asserts parser output against the test suites in test/
npm run build # rebuild the bundles in build/
npm run smoke # console.log the full parse of test/testData.json through the built bundleThe vitest suite pins existing parser output byte-for-byte, so any change to rendered markup is caught. See CHANGELOG.md for release history.
The parser does not escape HTML by design (except in code blocks) —
editor.js inline markup (<b>, <a>, <mark>…) inside block text is trusted
and passed through. Treat the editor.js JSON you feed it as trusted input: if
the data can be tampered with (e.g. stored user content from an untrusted
source), sanitize the resulting HTML at your render boundary with a
server-side sanitizer.
- 1.x → 2.0 (TypeScript rewrite): deep imports such as
require("editorjs-parser/build/Parser.node")are gone — import the package root instead (require("editorjs-parser")/import edjsParser from "editorjs-parser"), which works unchanged. Bundles moved frombuild/todist/; type declarations now ship in the package. - 2.x → 3.0: paragraphs no longer render with padding spaces
(
<p class="paragraph">text</p>). Feeding parsed HTML back into the editor no longer grows paragraphs. Want the old spacing back?new edjsParser(undefined, { paragraph: (data, config) => `<p class="${config.paragraph.pClass}"> ${data.text} </p>`, });
- 3.0 → 3.1: no breaking changes — block tunes, the
delimiter.tagconfig, and fixes foralt="undefined"and header-level clamping. - 3.1 → 3.2: no breaking changes —
alertblock, third-party alignment tune, and null table cells no longer render as literalnulltext (previously-broken output only).
Full history: CHANGELOG.md.
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change. Please make sure to update tests as appropriate.
For any issue or feature request, please open an issue.