UV: Modern Python Package Manager That's 10-100x Faster

Managing Python dependencies has always been a challenge. We need to install pip, virtualenv, pyenv, and poetry separately, then configure each tool individually. UV arrives as an all-in-one solution that replaces all these tools with 10-100x faster performance. Written in Rust by the Astral team (creators of Ruff), UV is a next-generation Python package manager that transforms how we work with the Python ecosystem.

In this tutorial, we will learn UV comprehensively from installation, project management, virtual environments, to a practical example of building an ML project from scratch.

Why UV?

Before diving into the technical details, let's understand why UV is worth considering:

• Exceptional speed: 10-100x faster than pip thanks to Rust implementation and global caching
• All-in-one: Replaces pip, pip-tools, virtualenv, pyenv, and poetry in a single binary
• Built-in lockfile: Reproducible builds without additional configuration
• Python version management: Install and switch Python versions without pyenv
• Cross-platform: Works consistently on Linux, macOS, and Windows
• Drop-in replacement: Compatible with the existing pip ecosystem

Installing UV

Linux and macOS

curl -LsSf https://astral.sh/uv/install.sh | sh

Windows (PowerShell)

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Using pip (alternative)

pip install uv

Using Homebrew (macOS)

brew install uv

Verify the installation:

uv --version

Updating UV

uv self update

Creating Projects with uv init

UV provides an easy way to start new Python projects:

uv init my-project
cd my-project

This command creates the following project structure:

my-project/
├── .python-version
├── README.md
├── hello.py
└── pyproject.toml

The generated pyproject.toml:

[project]
name = "my-project"
version = "0.1.0"
description = "Add your description here"
readme = "README.md"
requires-python = ">=3.12"
dependencies = []

Initialization Options

# Create project as a library (with src layout)
uv init --lib my-library

Create project as an application
uv init --app my-app

Specify minimum Python version
uv init --python 3.11 my-project

Create project in current directory
uv init

For library projects, the structure will be different:

my-library/
├── .python-version
├── README.md
├── pyproject.toml
└── src/
    └── mylibrary/

        ├── init.py
        └── py.typed

# Add a package
uv add requests

Add with specific version constraints
uv add "pandas>=2.0,<3.0"

Add multiple packages at once
uv add numpy scikit-learn matplotlib

Add development dependency
uv add --dev pytest ruff mypy

Add optional dependency group
uv add --group docs sphinx sphinx-rtd-theme

# Remove a package
uv remove requests

Remove a dev dependency
uv remove --dev mypy

uv tree

my-project v0.1.0
├── numpy v1.26.4
├── pandas v2.2.1
│   ├── numpy v1.26.4
│   ├── python-dateutil v2.9.0
│   │   └── six v1.16.0
│   ├── pytz v2024.1
│   └── tzdata v2024.1
└── scikit-learn v1.4.1
    ├── joblib v1.3.2
    ├── numpy v1.26.4
    ├── scipy v1.12.0