Skip to content

pixi global install#

Installs the defined packages in a globally accessible location and exposes their command line applications.

Need to specify build strings or hardware-specific packages?

For advanced package specifications including build strings, see the Package Specifications guide.

Examples (both syntaxes work):

# Equals syntax (compact)
pixi global install "pytorch=*=*cuda*" --channel pytorch
pixi global install "jax=*=*cuda*"

# Bracket syntax (explicit)
pixi global install "pytorch [build='*cuda*']" --channel pytorch
pixi global install "jax [build='*cuda*']"

Tip

Running osx-64 on Apple Silicon will install the Intel binary but run it using Rosetta

pixi global install --platform osx-64 ruff

Note

When you pass --path with a local .conda archive, Pixi copies the file into PIXI_HOME/conda-files and installs from that managed copy. Supplying any other kind of path keeps the original location unchanged.

Trusted virtual package detectors supply host capabilities for native-platform solves. For a foreign --platform, use CONDA_OVERRIDE_<NAME> to supply detector-defined virtual packages.

After using global install, you can use the package you installed anywhere on your system.

Usage#

pixi global install [OPTIONS] [PACKAGE]...

Arguments#

  • <PACKAGE>
    The dependency as names, conda MatchSpecs
    May be provided more than once.

Options#

  • --path <PATH>
    The path to the local package
  • --build-backend <BUILD_BACKEND>
    The build backend to build the source with, when the source does not provide its own package manifest (or to override the one it has). Accepts a name with an optional version constraint, e.g. pixi-build-rust or "pixi-build-rust>=0.3,<0.4"
  • --package <KEY=VALUE>
    Additional fields of the inline package definition, as DOTTED_KEY=TOML_VALUE pairs that are recorded under the package key of the dependency, e.g. host-dependencies.hatchling="*" or build.config.extra-args=["--all-features"]
    May be provided more than once.
  • --channel (-c) <CHANNEL>
    The channels to consider as a name or a url. Multiple channels can be specified by using this field multiple times
    May be provided more than once.
  • --platform (-p) <PLATFORM>
    The platform to install the packages for
  • --environment (-e) <ENVIRONMENT>
    Ensures that all packages will be installed in the same environment
  • --expose <EXPOSE>
    Add one or more mapping which describe which executables are exposed. The syntax is exposed_name=executable_name, so for example python3.10=python. Alternatively, you can input only an executable_name and executable_name=executable_name is assumed
    May be provided more than once.
  • --with <WITH>
    Add additional dependencies to the environment. Their executables will not be exposed
    May be provided more than once.
  • --force-reinstall
    Specifies that the environment should be reinstalled
  • --no-shortcuts
    Specifies that no shortcuts should be created for the installed packages

Config Options#

Git Options#

Description#

Installs the defined packages in a globally accessible location and exposes their command line applications.

Example:

  • pixi global install starship nushell ripgrep bat
  • pixi global install jupyter --with polars
  • pixi global install --expose python3.8=python python=3.8
  • pixi global install --environment science --expose jupyter --expose ipython jupyter ipython polars

Examples#

pixi global install ruff
# Multiple packages can be installed at once
pixi global install starship rattler-build
# Specify the channel(s)
pixi global install --channel conda-forge --channel bioconda trackplot

# Support full conda matchspec
pixi global install python=3.9.*
pixi global install "python [version='3.11.0', build_number=1]"
pixi global install "python [version='3.11.0', build=he550d4f_1_cpython]"
pixi global install python=3.11.0=h10a6764_1_cpython

# Install for a specific platform, only useful on osx-arm64
pixi global install --platform osx-64 ruff

# Install a package with all its executables exposed, together with additional packages that don't expose anything
pixi global install ipython --with numpy --with scipy

# Install into a specific environment name and expose all executables
pixi global install --environment data-science ipython jupyterlab numpy matplotlib

# Expose the binary under a different name
pixi global install --expose "py39=python3.9" "python=3.9.*"