pipx

Install application in isolated environment

$ pipx install [httpie]

$ pipx run [black] [file.py]

$ pipx list

$ pipx upgrade [youtube-dl]

$ pipx upgrade-all

$ pipx uninstall [ansible]

$ pipx inject [ipython] [matplotlib]

$ pipx reinstall-all

pipx command [options] [package]

pipx installs Python command-line applications in isolated virtual environments. Each application gets its own environment, preventing dependency conflicts while keeping executables available globally.
Unlike pip install, pipx creates separate environments for each tool. This means applications like black, flake8, and ansible can coexist even if they require conflicting package versions.
The run command executes applications without permanent installation. It downloads, runs, and cleans up - perfect for one-time use or trying new tools. Frequently run packages are cached for speed.
Inject adds extra packages to an application's environment. This is useful when an application has plugins or optional dependencies that need to be installed together.
The ensurepath command configures shell PATH to include pipx binary locations. This is typically needed once after initial pipx installation.

install PACKAGE

Install package in isolated environment.

Run app in temporary environment.

Uninstall package.

Upgrade package.

Upgrade all installed packages.

List installed applications.

Add packages to app's environment.

Remove injected packages.

Reinstall package.

Reinstall all packages.

Add pipx directories to PATH.

Include dependencies' executables.

Python interpreter to use.

Include system packages.

Arguments to pass to pip.

Force installation.

Verbose output.

Only for command-line applications, not libraries. Each app uses disk space for its environment. Injected packages may cause conflicts. Python version is fixed at install time. Some apps may need system dependencies.

pipx was created by Chad Smith around 2018 to address the problem of installing Python CLI tools cleanly. It was inspired by npx from the Node.js ecosystem. The project became a PyPA recommended tool for installing applications, while pip remains recommended for libraries.

pip(1), pipenv(1), poetry(1), virtualenv(1)