Reproducible Dependency Management with Pixi

Learn how to manage your conda- and pip-based Python dependencies with Pixi. With its declarative approach, Pixi provides a reliable way to create reproducible Python environments. Its design is based on experiences drawn from tools such as pip, conda, and mamba, as well as Rust's cargo.

This hands-on tutorial teaches you how to use Pixi for managing Python dependencies in a reproducible way. Pixi is a modern package manager that bridges the gap between the conda and pip ecosystems, offering the best of both worlds.

There are two main approaches to Python dependency management:

• pip-based: Uses PyPI packages
• conda-based: Uses conda packages from channels like conda-forge

Pixi offers a hybrid approach that works with both package types. Its design is inspired by Rust's cargo, providing a unified, declarative interface for managing dependencies. Beyond Python packages, Pixi can also manage software written in other languages.

What You Will Learn

• How to set up reproducible Python environments with Pixi
• How to manage both conda and pip dependencies in a single project
• How to create multi-platform projects
• How to define tasks and workflows in pixi.toml
• How to work with multiple environments (testing, docs, different Python versions)
• How to build and package Python projects with Pixi
• Best practices for using Pixi in real-world projects

Audience

This tutorial is designed for participants who:

• Have solid Python knowledge
• Are familiar with basic use of pip or similar tools
• Would like to learn what Pixi brings to dependency management
• Would like to get hands-on experience with Pixi

Prerequisites

• Comfortable writing Python code
• Basic familiarity with package installation (using pip or conda)
• Understanding of virtual environments (helpful but not required)
• Basic command-line usage

No prior experience with Pixi is required. Familiarity with poetry, conda, or other dependency management tools is helpful but not necessary.

Please install Pixi before the tutorial.

Your trainer: Dr. Mike Müller

Dr. Mike Müller has been teaching Python professionally since 2004 — with over 580 courses and 1,500 teaching days under his belt. A fixture at Python conferences worldwide, he has delivered 75+ tutorials at events including PyCon US (29 tutorials since 2007), EuroPython, and PyData. As a senior trainer at Python Academy, Mike combines deep technical expertise with a hands-on teaching style: live coding from scratch, progressive complexity, and real-world examples that participants can apply immediately.