Converting Markdown to HTML with emem

March 12, 2017
Updated: March 16, 2017

I have always wanted a way to create HTML documents from my Markdown files. Initially, I simply wanted to have HTML files from my text files so that I can view them nicely on my phone. Later, I also wanted a way to create these files so that I can upload them on the internet and view them on other devices. There are tools that exist that does just that. I tried them, then I found myself frequently tweaking the output just to make it look acceptable. None of them fitted my criteria: easy to build, easy to use, and produces decent output.

I wrote emem as a response to those needs. Emem is a small utility that takes in Markdown input either from stdin or disk file, then it produces a HTML output that is decent enough, at least, for regular viewing.

Table of contents


Emem is available via Nix. If you don’t have Nix yet, you may install it with:

$ curl | bash

You may then install emem with:

$ nix-env -i emem

When you’re done installing it, you may check the version number with:

$ emem --version

The latest version is 0.2.43.


At the most basic level, simply running emem against a Markdown file produces a basic, yet complete HTML file with all the necessary resources for correct page display. Applying emem on a file named

$ emem

will create the following tree:


While, running emem like so:

$ emem -s

will create the following tree:


The -s option, removes the need to create a separate resource directory, and stuffs all the resources needed into the output file making it easy and conducive to view the output documents on devices like phones and tablets.

Take note that the document title that will be used is the basename of the input file. So, for, it will generate <title></title> in the head tag. If you format your Markdown files, such that the first two lines look like:

Foo Bar

the first line will serve as the document title. To do so, run:

$ emem -F

resulting to <title>Foo Bar</title>.

That’s all nice and dandy, but if you want to create just a rudimentary document without all the bells and whistles, use the plain mode:

$ emem -Rp

The -R option instructs emem not to build the resource files, while the -p disables CSS and JavaScript.

If you want to change the name of the output file, use -o:

$ emem -o my-file.html

If you have Markdown files in ~/Desktop, you can convert them all to HTML in one fell swoop:

$ emem ~/Desktop

If you don’t like the default page width—40 em—use -f to use the available browser page display width:

$ emem -f

A feature that I like a lot is merging. This allows one to combine multiple files into a single output. For example, if you have,, and, you can merge them together to index.html with -m:

$ emem -mo index.html

If you are publishing for the web, it is imperative that you specify values for the description and keyword meta attributes. You may do so with the -D and -K options, respectively:

$ emem -D 'A diary about lobsters and crabs' \
-K 'lobsters, crabs, blog, journal, sea foods, monsters'

It is also possible to insert arbitrary content in the head tag with the -H option. For example, to specify OGP values:

To use Open Graph Protocol values:

$ emem -D Meh -K 'foo, bar, baz' \
--og-title "Crabs and Lobsters" --og-type article \
--og-url "" --og-image "" \

If you want to use Google Analytics, specify your 9-digit code, including the hyphen::

$ emem -D Foo -K 'qux, quux' -A 12345678-9

If the contents of the site is predominantly not in English, it is good to specify what language it is, to help search engines index your site properly; it also helps software, like screen readers, determine what language to use for the speech. For this, use the -l option:

$ emem -D 'Kie estas ĝin' -K 'kukurboj, hundegoj, afiŝoj' -l eo

A complete list of the supported languages of modern browsers can be found here.

There are times when I don’t want to break the edit loop when working with the input files, and I just want the HTML files to be created whenever there are new changes to the source Markdown files. In that scenario, I invoke the continuous build mode, with the -c option:

$ emem -c

The other options can be mixed with -c to fine-tune the output. For example, to build plain output in continuous mode:

$ emem -Rpc

Emem checks for changes to every 200 ms. If a change was detected, it will rebuild README.html. The timeout between checks can be changed with the -t option. To specify a 1 min timeout:

$ emem -Rpc -t 1000

To view all the supported options:

$ emem --help

Closing remarks

For this whole journal, I’m able to get a 90+ score from Google PageSpeed Insights, a Mobile-Friendly rating from Google Mobile-Friendly Test, and a performance grade of B from Pingdom. If you can control the web server parameters, you may even get a performance grade of A, when you leverage browser caching, and specify the Vary: Accept-Encoding header. I’m using GitHub Pages, so it’s a different story for me.

I’m quite happy with the output that emem produces. It’s fast enough and I can extend it easily. I even use it for my personal and work documentation. I also use it with Emacs to create a web previews of Markdown buffers using shell-command and w3m. To see emem in actual usage, go here.

If you know a bit of Clojure, fork it and hack away!