The first steps are to install docker, which is used to run jobs inside an isolated environment known as a “container”, and the Python package pycon which provides tools to submit, run, and manage batch jobs within a specified peer network.

Install docker

The python compute network uses docker to run containerized jobs on each worker node. At the time of this writing, only Mac and Linux operating systems are supported by the network.


The fastest and easiest way to get docker on macOS is to install the Docker Desktop application.


macOS must be version 10.13 or newer, and hardware must be 2010 or newer.


Download the Docker Desktop installer, double click Docker.dmg to open the installer then drag and drop the docker icon onto the applications folder. See the official docker for mac install instructions for details.

You should see a docker icon appear in your menu bar upon starting the Docker Desktop application. You’ll also be able to query the docker CLI, e.g. check the output of docker --help, to verify a successful installation.


There is no single method for installing docker on Linux; the installation instructions are specific to each distribution. The official docker documentation includes instructions for CentOS, Debian, Ubuntu, and Fedora. However, docker is ubiquitous enough that official instructions can be found for just about every Linux distro, e.g. to install docker on Arch Linux follow the instructions on the Arch Wiki.

Install pycon package

To interact with the compute network, you’ll need to install the pycon python package which is used to submit, run, and manage batch jobs. The package requires python3 and is installable with the python package manager pip:

pip3 install pycon

If the package is correctly installed, you should be able to call pycon-submit --help and see basic help output.

Per usual, you can also clone and install the package locally if you want access to the latest development version, i.e.

git clone
cd python_compute_network
pip3 install .