Writing and consuming plugins for Phenomic

For now, Phenomic supports plugins via the phenomicLoader.

You can easily do almost anything you want with a plugin to affect a file consumed by the phenomicLoader.

A plugin is just a function that accepts some parameters and that must return a modified collection item that will be added to the collection.

The idea is to work on this kind of object:

  // front-matter data
  head: {
    title: "Phenomic Is Using React, You'll Never Guess What Happened Next!",
    date: "2016-12-31",
    clickbait: true,
  // content of the file below the front-matter
  body: "React means the website is not static, but dynamic! Blah blah [...]",

Plugin might be used to transform the content depending on its format( Markdown, AsciiDoc, Textile. Txt2tags LaTeX.)

See the last section to know how to write a plugin (spoiler: it's easy).

#Existing Plugins

For now, plugins are accessible in phenomic/lib/loader-plugin-*.

import initHeadPropertyFromConfig from "phenomic/lib/loader-plugin-init-head-property-from-config"

This might change if phenomic is split into multiple packages. See #598 for more informations.


This plugin initializes the body property from data retrieved in the file. It takes the content of the input that is below the front-matter.


This plugin initializes in the head property from defaultConfig in the webpack phenomic configuration section. It won't override existing key/values in the head if there is any.


This plugin initializes in the head property from data retrieved in the file. It takes the front-matter of the input and map it as key => value.


This plugin initializes in a raw property. This property contains the entire file as raw data. Useful if you front-end need to handle the content of the file.


This plugin initializes in a rawBody property. This property contains the content of the file that is below the front-matter, as raw data. Useful if you front-end need to handle the content of the file.


This plugin initializes a description property in the head, based on the content accessible below the front-matter. This plugin assumes your content is markdown and will try to strip the markdown in order to offer raw text, handy for meta description, content preview etc. By default will strip to 140 characters (but it won't cut into the middle of a word). You can pass options to phenomic section in webpack configuration.

  pruneLength: 140, // max length of the description
  pruneString: "…", // string to append after the description


This plugin will transform the body property into html. This plugin will assumes your content is markdown and will use remark with some plugins for the transformation.


A preset is just an array of plugins. It's handy to share a collection of plugins. To use a preset, just spread it into the plugins options:

// ..
    plugins: [

⚠️ If you don't know what spread operator is (...), you can see it as a way to concat to arrays. You can write this as plugins: myPreset.concat(anotherPlugin) but this notation is more verbose, especially if you have multiple plugins.

#Existing presets

Phenomic provides the following presets:


  • phenomic/lib/loader-plugin-init-head-property-from-config
  • phenomic/lib/loader-plugin-init-head-property-from-content
  • phenomic/lib/loader-plugin-init-body-property-from-content

🛠 This preset is kind of the phenomic default requirement. Use it if you want to use classic files with a front-matter and any text format. Feel free to take a look at markdown preset to implement your own engine!


  • phenomic/lib/loader-preset-default
  • phenomic/lib/loader-plugin-markdown-init-head.description-property-from-content
  • phenomic/lib/loader-plugin-markdown-transform-body-property-to-html

❤️ This preset is the one used by default in Phenomic. It allows you to consume common markdown files that have a front-matter out of the box.

#Consuming plugins

Let say you want to use something else than the default markdown transformation.

You can use some plugins, but not the one that will transform the markdown as HTML.

import { phenomicLoader } from "phenomic"

  // ...

  // webpack 1
  // the problem with this approach is that it limit one instance of phenomicLoader
  // This won't work with webpack 2 since configuration is validated
  // webpack 2 accepts this arguments under `query` loader parameter
  phenomic: {
    plugins: [
      // here are the unopininated default plugins

      // Instead of specifing via the preset, you can cherry pick some,
      // require("phenomic/lib/loader-plugin-init-head-property-from-config").default,
      // require("phenomic/lib/loader-plugin-init-head-property-from-content").default,
      // require("phenomic/lib/loader-plugin-init-body-property-from-content").default,

      // ...require("phenomic/lib/loader-preset-markdown").default
      // The commented preset above is part of the default renderer.
      // You can also cherry pick on plugin or the other
      // require("phenomic/lib/loader-plugin-markdown-init-head.description-property-from-content").default,
      // require("phenomic/lib/loader-plugin-transform-md-body-property-to-html").default,

      // here is an example of another transformation
      ({ result }) => {
        return {
          body: doYourThing(result.body),

          // doYourThing() can be any transformation you want.
          // Per previous plugins, ``body`` will contain the raw input
          // passed to the loader, with the front-matter removed.
          // Feel free to take a look at the plugins source before,
          // it's very short !
          // https://github.com/MoOx/phenomic/blob/master/src/

  module: {
    // webpack 1
    loaders: [
    // webpack 2
    rules: [
        test: /\.(md|markdown)$/,
        loader: phenomicLoader,
        query: {
          context: path.join(config.cwd, config.source),

          // only webpack 2 can accept all phenomicLoader option here
          // since webpack 1 does not allow anything that can't be serialized to json

      // ...

#Writing plugins

A plugin is just a function that accepts the following parameters:

  frontMatter: GrayMatterResult,
  result: PhenomicCollectionItem,
  options: PhenomicLoaderOptions,
}) => PhenomicCollectionItem

▶️ frontMatter is the result of the parsing of the file via gray-matter package and is an object:

  orig: string, // original content (the file)
  data: Object, // key => value of the front-matter content
  content: string, // the raw content below the front-matter

▶️ result is a Phenomic collection item. In the first plugin it just an empty object, but at the end of the plugin pipeline it should offer the following object

  head: Object, // key => value of the front-matter content
  body: string, // html that will be injected in Phenomic ``<BodyContainer>``

  // Keys below are reserved for Phenomic (and added after the plugins execution)

  // ⚠️ Note that the following keys are added by the phenomicLoader itself
  // you cannot affect those values.
  __filename: string, // original relative filename (eg: usage/plugins.md)
  __url: string, // short uri (eg: usage/plugins/)
  __resourceUrl: string, // long uri (eg: usage/plugins/index.html)
  __dataUrl: string, // data uri (for optimized navigation, eg: usage/plugins.{hash}.json)

▶️ options are the options of the current loader definition.

It's currently a merge of webpack phenomicLoader query options & phenomic.phenomicLoader section of your webpack configuration. ⚠️ Webpack 2 won't accept the global phenomic object.

The plugin must return an result item, that will be added to Phenomic collection.
You can add any key you want, just be sure to provide the head object and the body.

#Example of a plugin

The plugin below is a plugin that transform the body (in this case, set before in another plugin).

const transformMdBodyPropertyToHTML = ({ result }) => {
  return {
    ...result, // spread the entire result
    // and override the body property by applying a transformation
    body: doYourThingToHTML(result.body),

⚠️ If you want to see some examples for plugins, take a look at