Main Features

User Interaction

On the right side of any page, you’ll notice links to edit the page, or open an issue. This ensures that any time you have a question or want to suggest or request a change, you can do so immediately and link directly to the section of interest. The sections on the page also have permalinks so you can link directly to them.

The entire site, including posts and documentation, is indexed and then available for search at the top or side of the page. Give it a try! The content is rendered into window data that is used by lunr.js to generate the search results. If you want to exclude any file from search, add this to its front end matter:

layout: null
excluded_in_search: true

The example above is for a javascript file in the assets folder that is used as a template, but should not be included in search.

If you have an external site with a search GET endpoint (meaning one that ends in ?q=<term>, then you can automatically link page tags to search this endpoint. For example, on an HPC site I’d want a tag like “mpi” to do a search on for mpi. See the tags section below for how to configure this.


Documentation pages should be written in the docs folder of the repository, and you are allowed to use whatever level of nesting (subfolders) that works for you! It’s a Jekyll collection, which means that you can add other content (images, scripts) and it will be included for linking to.


The url that will render is based on the path. For example, if we had the following structure:


The first page (akin to the one you are reading) would render at it’s path, /docs/getting-started/.


From that page, we could provide the direct path in markdown to any subfolder to link to it, such as the second getting started page for sherlock:


Here is an example link to a relative path of a file ( in the same directory, and from that page you can test linking to a subfoldr. In the case of not having a subfolder, we could write the link out directly:

[example]({{ site.baseurl }}/docs/clusters/sherlock/

or just put the relative path:


or better, there is a shortand trick! We can use the provided “includes” template to do the same based on the path to create a link:

{% include doc.html name="Sherlock Cluster" path="clusters/sherlock/getting-started" %}

The path should be relative to the docs folder.


The pages folder uses the same page layout, but is not part of the docs collection. The two are provided to create a distinction between website pages (e.g., about, feed.xml) and documentation pages.

Whether you place your page under “pages” or “docs,” for those pages that you want added to the navigation, you should add them to _data/toc.yml. If you’ve defined a permalink in the front end matter, you can use that (e.g., “About” below). If you haven’t and want to link to docs, the url is the path starting with the docs folder. Here is an example (currently the active example):

- title: Documentation
  url: docs
    - title: "Getting Started"
      url: "docs/getting-started"
        - title: Features
          url: "docs/getting-started#getting-started"
        - title: Development
          url: "docs/getting-started#development"
        - title: Customization
          url: "docs/getting-started#customization"
    - title: "Extras"
      url: "docs/extras"
        - title: Quizzes
          url: "docs/extras/example-quiz"
    - title: "About"
      url: "about"
    - title: "News"
      url: "news

If you want to add an external url for a parent or child, do this:

  - title: GitHub Repository

News Posts

It might be the case that your site or group has news items that would warrent sharing with the community, and should be available as a feed. For this reason, you can write traditional posts in the _posts folder that will parse into the site feed The bottom of the page links the user to a post archive, where posts are organized according to the year.


Buttons come in a nice array of colors. Here is the code for a basic example, and you’d want to vary the .btn-<tag> to get different classes.

<button class="btn btn-success">.btn-success</button>


For news post items, it’s nice to be able to tag it with something that indicates a status, such as “warning” or “alert.” For this reason, you can add badges to the front end matter of any post page, and they will render colored by a type, with the tag of your choice. For example, here is an example header for a post:

title:  "Two Thousand Nineteen"
date:   2019-06-28 18:52:21
categories: jekyll update
 - type: warning
   tag: warning-badge
 - type: danger
   tag: danger-badge

And here is the post preview with the rendered badges that it produces:

warning-badge danger-badge

And the other badges that you can define include success, info, secondary, and primary.

success-badge info-badge secondary-badge primary-badge


{% include alert.html type="info" title="Here is another!" %}

Just for fun, here are all the types:


You can include block quotes to emphasize text.

Here is an example. Isn’t this much more prominent to the user?


Initially (on OS X), you will need to setup Brew which is a package manager for OS X and Git. To install Brew and Git, run the following commands:

/usr/bin/ruby -e "$(curl -fsSL"
brew install git

If you are on Debian/Ubuntu, then you can easily install git with apt-get

apt-get update && apt-get install -y git

Install Jekyll

You can also install Jekyll with brew.

$ brew install ruby
$ gem install jekyll
$ gem install bundler
$ bundle install

On Ubuntu I do a different method:

git clone ~/.rbenv/plugins/ruby-build
echo 'export PATH="$HOME/.rbenv/plugins/ruby-build/bin:$PATH"' >> ~/.bashrc
exec $SHELL
rbenv install 2.3.1
rbenv global 2.3.1
gem install bundler
rbenv rehash
ruby -v

# Rails
curl -sL | sudo -E bash -
sudo apt-get install -y nodejs
gem install rails -v 4.2.6
rbenv rehash

# Jekyll
gem install jekyll
gem install github-pages
gem install jekyll-sass-converter

rbenv rehash

Get the code

You should first fork the repository to your GitHub organization or username, and then clone it.

$ git clone<username>/mkdocs-jekyll.git docs
$ cd docs

You can clone the repository right to where you want to host the docs:

$ git clone<username>/mkdocs-jekyll.git docs
$ cd docs


Depending on how you installed jekyll:

jekyll serve
# or
bundle exec jekyll serve


We provide a CircleCI configuration recipe that you can use to preview your site on CircleCI before merging into master. You should follow the instructions to set up a project, and then in the project settings be sure to enable building forked build requests, and to cancel redundant builds. The preview will be built on CircleCI, and saved to static files for you to browse. The only change you will need is to edit the static files location to be the name of your respository, which is at te bottom of the .circleci/config.yml file:

      - store_artifacts:
          path: ~/repo/_site
          destination: mkdocs-jekyll

In the above, the destination should coincide with your repository name. Remember that for most links, CircleCI won’t honor an index.html file in a subfolder (e.g., subfolder/index.html will not be served as subfolder/, so for example, you might need to turn this:


into this:




To edit configuration values, customize the _config.yml. Most are documented there, and please open an issue if you have questions.

Adding pages

To add pages, write them into the pages folder. You define urls based on the permalink attribute in your pages, and then add them to the navigation by adding to the content of _data/toc.yml.


If you include tags on a page, by default they will link to the tags page on the site. However, if you define a tag_search_endpoint url in your configuration file, by clicking the tag, the user will be taken to this page to search for it. As an example, we define the current search endpoint to be Ask Cyberinfrastructure, and page tags link to a search on it:

tag_color: danger # danger, success, warning, primary, secondary, info

Note that you can also choose a color! The tags appear at the top of the page, as they do on this page. The tags should be defined like this in the front end matter:

 - jekyll
 - github

They are appended to the first h1 block, so generally your pages should have a header. If you comment out this variable, then each of your tags will link to it’s appropriate spot on the tags page linked above.

tag_color: primary # danger, success, warning, primary, info, secondary


As of version 0.0.12, mkdocs-jekyll has support for basic quizzes! These are intended to help educate your users about the content of your documentation. For a quiz, you can add a new file to the folder _data/quizzes, and write a questions file based on the format shown in _data/quizzes/example-quiz.yml. Here is a simple example of a multiple choice question (which can also serve as True/False):

title: This is the Quiz Title
randomized: false

 - type: "multiple-choice"
   question: "True or False, Pittsburgh is West of Philadelphia"
    - choice: True
      correct: true
    - choice: False
      correct: false
   followup: | 
      The answer is True! Pittsburgh is 304.9 miles West of 
      Philadelphia, or approximately a car ride of 
      4 hours and 52 minutes. Buckle up!

The quiz is rendered with a “Show Answer” button below each question, and when the user clicks it, any questions that are flagged with correct: true will be bolded, and if a followup section is included, it will be displayed. See the live example at the end of this page.



If you include a title, it will be rendered at the top of the quiz. This is optional - you can leave it out and add it before the include on the page.


If you want your questions to be presented randomly, just add randomized: true to the data.

Example Quiz

If I want to include the quiz located at _data/quizzes/example-quiz.yml, I can do so like this:

{% include quiz.html file='example-quiz' %}

The rendered quiz is shown here:

This is the Quiz Title

What is your favorite color?

1. Red
2. Blue
3. Green

True or False, Pittsburgh is West of Philadelphia

1. true
2. false

A Nested Page

Let’s say we want a page that doesn’t have a permalink defined, and is not included in the table of contents (_data/toc.yml). This means that it will render based on it’s path. For example if the page is in docs/, the url will be docs/example-page/.

Now let’s say we want to link to a subfolder, specifically with this setup:

docs/  (-- we are here
  subfolder/  (-- we want to link here

You can provide the relative path to the file, like subfolder/ and Jekyll will handle parsing it. For example:

And here is the same link, but generated with the include statement:

{% include doc.html name="here" path="subfolder/example-page" %}


The news section lists all the articles inside the _posts folder.

It can be referred through the link: