Dependencies For full flexibility, Cubonacci lets users define exactly which packages are required for running all the different steps in the process. By placing a requirements.txt in the root of the repository, Cubonacci will know to install these custom dependencies in the environment where your code will run. This file follows the same conventions as is common in the Python ecosystem. Structure Every line represents a singular package that is required. Preferably this includes the version of the package so that breaking changes in newer versions will not introduce unforeseen errors in your code. Specifying versions can be done by adding == and the version number behind the name. Without additional information, these packages are installed from PyPI, the Python Package Index. Installing packages from a git repository is also possible, this is done using a git URL with a prefix that starts with git. More information can be found here. Here is an example requirements.txt for a Python project in Cubonacci. As you can see, comments are possible by using #. numpy # Adding a version is highly recommended pandas==0.25.3 git+https://github.com/django/django.git@45dfb3641aa4d9828a7c5448d11aa67c7cbd7966 Private packages Referencing private packages either via Git or private PyPI servers is also possible, please send a message to support to set this up because this depends on the specific setup.