Skip to content


Subversion checkout URL

You can clone with
Download ZIP
Deploying should be easy. [currently inactive]
Python CSS JavaScript


depl - deploying should be easy

Travis-CI build status Coverage Status Number of PyPI downloads Latest PyPI version

depl is a pre-alpha prototype. We want to start a discussion about how developers deploy. Please tell us why you think depl is not the right tool for you!

Deploying stuff is hard, managing nginx and postgres painful. Why don't we solve that problem? Simple deployments should not involve tweaking configurations that you have never dealt with before. depl solves this.

Let's deploy Django to a live server! Create a file called depl.yml:

    - django:
        settings: settings_prod
            key: /path/to/key
            certificate: /path/to/cert

And run depl deploy in that same directory. After this two step process your website will be running on and It only works on Ubuntu servers for now (maybe also Debian), but will work on other systems in the future as well.

depl would connect to by SSH, install all necessary software packages and initialize your project. In our Django example this would involve installing nginx, uwsgi, postgresql (if required by Django's It would install all Python dependencies and run on your production settings.

Currently this is also working for Meteor and WSGI (which includes flask).

Why depl and not ansible, puppet, chef, docker or vagrant?

Have you ever tried those tools? ansible, puppet and chef are really complicated to understand. They are made for big deployments, for large scale operation engineers. You can read the blog posts below, they explain a bit more. Generally depl solves the use case for all small deployments on your own server, those tools do something different.

Docker and Vagrant are completely different tools. They make controlling VMs easy. They do not install your software. However, depl could be used in combination with either of them.

Blog Posts talking about depl


If you want to know more in general, just read the documentation at!


Currently depl supports Ubuntu/Debian (apt-get), the others:

  • Arch Linux (pacman) could be implemented
  • Fedora/CentOS/RHEL (rpm) could be implemented
  • Windows/Apple: Discussions needed (but probably not?!).

Depl is written in Python. Python 3 support will be ready once fabric supports Python 3.

Additional Ideas

  • DNS configuration
  • Central logging
  • High Availability tools like HAProxy to make it easy not to go offline for a few seconds while deploying.
  • Could generate e.g. an ansible recipe with depl. Possible?
  • Write a "depl server" with different backends (e.g. amazon clouds) to automate the process of deploying and increasing servers.
  • Add support for automatic travis testing with depl files. Is that even possible?


Read to check how you can contribute! There's also a small developer documentation available.


Use a virtual machine for testing, please. Why?

Something went wrong with that request. Please try again.