Create simple websites fast, now.

Poole is an easy to use Markdown driven static website generator. You write the content of your pages in Markdown and Poole creates a nice and simple site with a navigation menu. You don't need to learn a template or preprocessing engine.

Though Poole is made for simple sites, it has basic support for content generation by embedding Python code in page source files. This is a dirty merge of content and logic but for simple sites it's a pragmatic way to get things done fast and easy -- if you know Python you're ready to start.

Check the list of sites built with Poole (and feel free to add yours).

Flattr this


You should know Markdown and optionally Python if you want to use Poole's dirty content generation capability.

Your system should have installed Python ≥ 2.5 and python-markdown. Poole has been tested on Linux but should also work on other Unix systems and Windows (in theory, report an issue if it fails).

Getting Started

Clone oder download (zip, tgz) the repository and then put to your PATH:

$ hg clone /some/where/poole
$ export PATH=$PATH:/some/where/poole

TIP: You might want to add the last command to your ~/.bashrc.

Create and build a site project:

$ mkdir /path/to/site/project
$ cd /path/to/site/project
$ --init
$ --build
$ --serve

Done. You've just created a website! Browse http://localhost:8080/ and watch the example pages which have been created during initialization.

To write your own pages, use the example pages in the input folder as a starting point.

Run --build whenever you've made some changes in the input folder.

How It Works

Poole takes files from a project's input directory and copies them to the output directory. In this process files ending with md, mkd, mdown or markdown get converted to HTML using the project's page.html as a skeleton.

Additionally Poole expands any macros used in a page. Don't care about that for now ..

When running --build in a Poole project, an input directory like this:

|- input
    |- news.mkd
    |- foo.mdown
    |- images
        |- bar.png

will result in an output folder like that:

|- output
    |- index.html
    |- news.html
    |- foo.html
    |- images
        |- bar.png

Page Layout

Every Poole page is based on the skeleton file page.html. Hence adjusting the site layout means adjusting page.html and extending or replacing its CCS file input/poole.css.

The only thing you should keep in page.html are the embedded {{__content__}} and {{__encoding__}} expressions. Below is an almost minimal page.html file. It does not look nice but it's a clean starting point to build your own layout from scratch.

Minimal page.html:

    <meta http-equiv="Content-Type" content="text/html; charset={{ __encoding__ }}" />
    {{ __content__ }}

It's easy to apply one of the numerous free CSS templates out there to a Poole site. For more information read this blog post with step-by-step instructions.

Content Generation

Poole allows you to embed Python code in your pages to generate content:


Here is normal text in *markdown* flavor.
print "hello poole"
Did you know? The sum of 2 and 2 is {{ 2 + 2 }}.

This example demonstrates 2 ways to embed Python code, either as statements or as expressions:

  1. Everything between {% and %} are statements and whatever is printed to stdout during their execution is going to be part of the final HTML page.
  2. Everything between {{ and }} are expressions and their evaluation is going to be part of the final page.

TIP: Instead of the outer curly brackets { and } you can also use <!-- and --> to prevent syntax highlighting markdown editors from getting confused by the Python code.

TIP: To print the code surrounding tags literally, simply escape the opening tag with a backslash.

Outsource complex or frequently used code

To keep embedded code short and compact or to reuse it in several pages, it can be outsourced into a file called in a project's root folder (where the page.html file is located). Every public attribute in is available within embedded Python code blocks:

from datetime import date
def today():
    return"%B %d, %Y")

author = "Popeye"


This site has been updated on {{ today() }} by {{ author }}.

Builtin macros

Builtin macros can be used from the macros module as well as from python code in your pages and templates (just as if they are defined within your

Currently, there is only one builtin macro available.


Replace the characters that are special within HTML (&, <, > and ") with their equivalent character entity (e.g., &amp;). This should be called whenever an arbitrary string is inserted into HTML (i.e. use {{ htmlspecialchars(variable) }} instead of {{ variable }}).

Note that " is not special in most HTML, only within attributes. However, since escaping it does not hurt within normal HTML, it is just escaped unconditionally.

Working with pages

Next to stuff defined in the objects page and pages are available in embedded Python code. The first one is a dictionary describing the page in which the code is embedded. The second one is a list of all pages in the project.

The following attributes are always set in a page dictionary:

  • title: The page's title, by default its filename without extension (setting alternatives is described in the next section).
  • fname: Path to the page's source file, e.g. /path/to/project/input/stuff/
  • url: The page's relative URL, e.g. for a source page input/stuff/ this is stuff/news.html.

The example page.html file in a freshly initialized site project uses a page's title attribute:

<div id="header">
     <h1>a poole site</h1>
     <h2>{{ page["title"] }}</h2>

TIP: All items in a page dictionary are exposed as attributes, i.e. page["foobar"] is identical to page.foobar. Dictionary access is useful if an item may not be set, e.g.: page.get("foobar", "...").

Setting page attributes

Page attributes can be set at the top of a page's source file, in Python's configuration file style. They are delimited from the page's content by a line with 3 or more dashes.

A page can be translated in various languages. The latter are delimited by the directive lang:somelang. The associated title variable title_somelang can be used in tandem.


title: Hot News
title_fr: Nouvelles
foobar: King Kong
Here are some news about {{ page.foobar }}.
Did I say {% print(page.foobar) %}?

lang: fr
Voici quelques nouvelles a propos de {{ page.foobar }}.
Ai-je bien dit {% print(page.foobar) %} ?

That way you can also set a page's title explicitly, instead of using the file name. Other useful attributes to set are description and keywords, which get used by the default page.html file to set HTML meta tags. Here it comes in handy to set default page attributes in the file:

page = { "description": "some stuff", "keywords": "stuff" }

That way you can safely use the description and keywords attributes without bothering if they are really defined in every page.

Accessing page objects in the macros module

The objects pages and page are also available within That means you can define them as dummys and reference them in Poole updates them when loading the macros module.

page = {} # you can also set defaults here, see previous section
pages = []

def something():
    # when executing this, the page and pages objects above are up-to-date
    print page["title"]

Options and paths

Similarly to page and pages the following objects are available within embedded Python code and within the macros module:

  • options: The command line arguments given to Poole as parsed by Python's optparse module. For instance the base URL can be retrieved by options.base_url.
  • project: Path to the project's root directory.
  • input: Path to the project's input directory.
  • output: Path to the project's output directory.

Character encodings

In case you use non-ASCII characters, check the encoding options of Poole. In most cases working with non-ASCII strings should work straight forward if the options are set properly (default is UTF-8).

However, be aware that page variables defined within page source files and derived from a page's file name internally are handled as Python unicode objects. That means if you want to refer to non-ASCII page variable names and values form within embedded Python code or from, make sure to use unicode strings to reference them.

Custom file converters

If you use LESS or CleverCSS you'll be happy about the possibility to define custom converters Poole applies to selected files in its building process. Custom converters may be defined in using a dictionary named 'converter' with file name patterns as keys and a converter function as well as a target file name extension as values:

converter = {
    r'\.ccss': (ccss_to_css, 'css'),

The converter function ccss_to_css must accept the source file name and the destination file name as arguments. The destination file name is a suggestion (the source filename mapped to the project's output directory with the extension given in the converter dictionary) - you are free to choose another one:

import clevercss

def ccss_to_css(src, dst):
    # when `src` is '/path/to/project/input/foo.ccss'
    # then `dst` is '/path/to/project/output/foo.css'
    ccss = open(src).read()
    css = clevercss.convert(ccss)
    open(dst, 'w').write(css)


You can do some pretty fancy and useful things with inlined Python code and the macros module, for instance generate a list of blog posts or create an RSS file. Check out the example recipes.


Please use the issue tracker.