Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time


A cool and easy static site builder.

Jinja templates. Markdown. Pygments source code higlighting. Blog-aware. Atom feeds. Live preview. Clean URLs. MIT license. Minimal design. Words. Dots.

Build Status

Designed for Simplicity

Icecake aims to be simple. It has a simple license. It does not support lots of features. For example, it does not support plugins, themes, different templating backends, or any of a long list of other features.

Icecake is less than 750 lines of python and still has everything you need to have a pleasant time writing and updating your personal website or blog.

I hope that this simplicity makes it easy for you to read the code and add any feature(s) you can't live without.

If you are already familiar with how static site builders work and you're comfortable with Jinja templates, you can use the following commands to get started:

pip install icecake
icecake init
icecake preview

If these topics are new to you, keep reading for more details!


pip install icecake

How do I get pip?

pip usually comes with Python. If it's missing, you can get it here.

Creating a Website

First, we'll setup a site:

icecake init mysite
cd mysite

It will look something like this:

├── content
│   ├── articles
│   │   └──
│   ├── articles.html
│   ├── atom.xml
│   └── index.html
├── layouts
│   ├── basic.html
│   └── markdown.html
├── output (empty for now)
└── static
    └── css
        ├── main.css
        └── syntax.css

The starter site includes a minimal theme and the articles folder will help you start blogging right away (if you want to do that).

Run icecake preview to view the site. The site will be automatically regenerated when you make changes.

Generating the Site

You can run icecake build to build your site. Icecake will generate each page and then exit (or error). If you get an error you can use icecake build --debug to get some more detailed information about what is happening.

Internally, there is a 3 step process for generating the site.

  1. Icecake reads all of the files under content and renders them. Markdown is converted to HTML and Jinja templates are evaluated.
  2. Icecake writes all of the rendered files into output. articles/ becomes articles/hello-world/index.html so you get nice URLs on any hosting platform.
  3. Files from static are copied as-is to output (using the same directory structure as the originals).

A page's URL is based on the filename, without the file extension. For example, articles/ becomes articles/hello-world/. There is a special exception for files named index.html or We usually don't want these to end up as e.g. articles/index/. If you do actually want "index" to be in the URL you can explicitly set this by specifying the slug.

When you're ready, you can use rsync or s3cmd or an FTP client to publish output to the web.

Editing Content

You can write content in either Markdown or HTML. Markdown files (idenfified by .md or .markdown) are automatically parsed and rendered using the markdown.html template. Source code blocks are highlighted using Pygments.

HTML files are evaluating using Jinja. Refer to the Jinja Template Designer documentation for details. We will also highlight some important Icecake-specific features related to templates below.

Markdown files are handled as a special case inside Icecake, so you can't mix Markdown and Jinja in the same file. However, you can customize the markdown template by editing markdown.html or by overriding the template metadata field and specifying a new template. See the Page Metadata section for more info.

Page Metadata

At the top of each file you can add some metadata. You should add this to all your markdown pages. Metadata is usually not needed for HTML pages.

title = Installing Ruby the Correct Way
date = 2013-03-27
tags = ruby bundler rbenv ruby-build
slug = installing-ruby
template = custom-markdown.html

Your content starts here!
  • title (required) The title of your page
  • date (required) The date your page is published (for display only)
  • tags (optional) Space-separated list of tags, which can be used to categorize your page.
  • slug (optional) This will be used instead of the filename in the URL
  • template (optional) This overrides the template (markdown.html by default) used to render the page

Some other metadata are generated for you automatically:

  • filepath Relative path of the file inside content, such as articles/
  • folder Just the folder part like articles/
  • ext The file extension (.html or .md for example)
  • url The path part of the URL, such as /articles/hello-world/

These metadata are important not just to make your page display correctly, but also to query other pages in your templates. You will do this to make a list of all your pages, for example.

Rendering Metadata

Whenever you are writing a template the current page's metadata are available via their names, so you can show the page title via {{ title }} or the publish date via {{ date }}. You can also write if statements that reference this information.

To show a list of tags for your page you can write something like this:

<h1>{{ title }}</h1>

{% for tag in tags %}
    <span class="tag">{{ tag }}</span>
{% endfor %}

When Is Metadata Required?

Metadata is only required if you reference it somewhere, such as using {{ title }} or when using site (see below). I recommend always using it on Markdown files and only using it on HTML if you have to. The default Markdown templates assume that you have provided titles, tags, and publish dates in your metadata and won't work properly if these are missing.

There are a lot of things you can accomplish via Jinja template variables so you don't necessarily need to use metadata make your site to show up the way you want.

Warning: if you try to query a page using site (below) and a piece of metadata is missing, your query will fail.

Site Helpers

Site helpers are special features available in templates that allow you to query all pages and tags across your site. For example, if you want to incldue a list of your 5 latest blog entries on your homepage, the site helpers can do this for you.

Querying Pages

You can search across the pages on your site using site.pages.

site.pages(option=value, ...)

With no arguments, site.pages includes all pages on your site. You can filter this list using the following filter options:

  • path Filter the list of pages based on the path under content/. The path string is compared using startswith() so cake will match cake/chocolate but not chocolate/cake.
  • tag Filter based on a tag. If a page has the tag you specified, it will be included.
  • order Sort the list of results by the specified field, like date or title. Use -date to reverse the order.
  • limit Limit the number of items you get back. The limit you specify can be higher that the number of actual results; you'll just get as many as are available.

You can combine these options much like SQL. They are evaluated in the order listed above, so a path filter is applied first, second tag, third order, and finally limit.

Warning: If you try to sort based on a metadata property that is not specified on every item, sort will fail! Icecake does not enforce that all of your pages have the same metadata so this is up to you. Use icecake build --debug if you're having trouble figuring out which file(s) are missing which field(s).

We'll show two examples of how to use this below.

List Articles Tagged "Blog"

You can use this to create a blog index page, for example:

{% for page in site.pages(tag="blog", order="title") %}
    <a href="{{ url }}">{{ title }}</a>
{% endfor %}

List 5 Recent Articles

{% for page in site.pages(path="articles/", order="-date", limit=5) %}
    <a href="{{ url }}">{{ title }}</a>
{% endfor %}

If you want to match a folder named articles/ but not a file named articles.html, make sure to include / at the end!

Generating an Atom Feed

You can use site.atom to create an Atom feed for specific pages on your site. The query behavior works exactly the same way as site.pages so please refer to that for details.

Unlike site.pages the atom feed is simply printed out -- you don't need to iterate over it.

    feed_title="My Blog",

Obviously you should fill in your name and site URL above. Note that most of these options are required in order for the feed to work correctly, but if you want to skip one you can specify None as the value. Also, you can use "https://yoursite"+url to automatically set the feed URL to point to the current page.

Listing Tags

You can use site.tags to list all of the tags in use on your site. You cannot currently query or filter the list of tags.

{% for tag in site.tags() %}
{% endfor %}

Don't confuse this with tags!

Questions? Problems? Suggestions?

Open an issue!