The simplest way to use MJML in Django templates.

• Django from 1.8 to 4.0
• requests from 2.20.0 (only if you are going to use API HTTP-server for rendering)
• mjml from 2.3 to 4.10.1

1. Install mjml.

See https://github.com/mjmlio/mjml#installation and https://mjml.io/documentation/#installation

2. Install django-mjml.

$ pip install django-mjml

If you want to use API HTTP-server you also need requests (at least version 2.20):

$ pip install django-mjml[requests]

To install development version use git+https://github.com/liminspace/django-mjml.git@master instead django-mjml.

3. Set up settings.py in your django project.

INSTALLED_APPS = (
  ...,
  'mjml',
)

Load mjml in your django template and use mjml tag that will compile MJML to HTML:

{% load mjml %}

{% mjml %}
    <mjml>
    <mj-body>
    <mj-container>
        <mj-section>
            <mj-column>
                <mj-text>Hello world!</mj-text>
            </mj-column>
        </mj-section>
    </mj-container>
    </mj-body>
    </mjml>
{% endmjml %}

There are three backend modes for compiling: cmd, tcpserver and httpserver.

This mode is very simple, slow and used by default.

MJML_BACKEND_MODE = 'cmd'
MJML_EXEC_CMD = 'mjml'

You can change MJML_EXEC_CMD and set path to executable mjml file, for example:

MJML_EXEC_CMD = '/home/user/node_modules/.bin/mjml'

Also you can pass addition cmd arguments, for example:

MJML_EXEC_CMD = ['node_modules/.bin/mjml', '--config.minify', 'true', '--config.validationLevel', 'strict']

Once you have a working installation, you can skip the sanity check on startup to speed things up:

MJML_CHECK_CMD_ON_STARTUP = False

This mode is faster than cmd but it needs run a separated server process which will render templates.

MJML_BACKEND_MODE = 'tcpserver'
MJML_TCPSERVERS = [
    ('127.0.0.1', 28101),  # host and port
]

You can set several servers and a random one will be used:

MJML_TCPSERVERS = [
    ('127.0.0.1', 28101),
    ('127.0.0.1', 28102),
    ('127.0.0.1', 28103),
]

You can run servers by commands:

# NODE_PATH=/home/user/node_modules node /home/user/.virtualenv/default/lib/python2.7/site-packages/mjml/node/tcpserver.js --port=28101 --host=127.0.0.1 --touchstop=/tmp/mjmltcpserver.stop

28101 - port, 127.0.0.1 - host, /tmp/mjmltcpserver.stop - file that will stop server after touch.

For daemonize server process you can use, for example, supervisor:

/etc/supervisor/conf.d/mjml.conf

[program:mjmltcpserver]
user=user
environment=NODE_PATH=/home/user/node_modules
command=node
    /home/user/.virtualenv/default/lib/python2.7/site-packages/mjml/node/tcpserver.js
    --port=28101 --host=127.0.0.1 --touchstop=/tmp/mjmltcpserver.stop --mjml.minify=true --mjml.validationLevel=strict
stdout_logfile=/home/user/project/var/log/supervisor/mjml.log
autostart=true
autorestart=true
redirect_stderr=true
stopwaitsecs=10
stopsignal=INT

Or you can use docker-compose:

services:
  mjml-1:
    image: liminspace/mjml-tcpserver:latest
    restart: always
    ports:
      - "28101:28101"

  mjml-2:
    image: liminspace/mjml-tcpserver:latest
    restart: always
    environment:
      HOST: "0.0.0.0"
      PORT: "28102"
      MJML_ARGS: "--mjml.minify=true --mjml.validationLevel=strict"
    expose:
      - "28102"
    ports:
      - "28102:28102"

You also can build your own tcpserver with other versions of MJML by using docker/mjml-tcpserver file and editing arguments.

don't forget to install requests to use this mode.

This mode is faster than cmd and similar to tcpserver but you can use official MJML API https://mjml.io/api or run your own HTTP-server (for example https://github.com/danihodovic/mjml-server) to render templates.

MJML_BACKEND_MODE = 'httpserver'
MJML_HTTPSERVERS = [
    {
        'URL': 'https://api.mjml.io/v1/render',  # official MJML API
        'HTTP_AUTH': ('<Application ID>', '<Secret Key>'),
    },
    {
        'URL': 'http://127.0.0.1:38101/v1/render',  # your own HTTP-server
    },
]

You can set one or more servers and a random one will be used.