Skip to content

Project Configuration

The pixi.toml is the pixi project configuration file, also known as the project manifest.

A toml file is structured in different tables. This document will explain the usage of the different tables. For more technical documentation check pixi on


We also support the pyproject.toml file. It has the same structure as the pixi.toml file. except that you need to prepend the tables with tool.pixi instead of just the table name. For example, the [project] table becomes [tool.pixi.project]. There are also some small extras that are available in the pyproject.toml file, checkout the pyproject.toml documentation for more information.

The project table#

The minimally required information in the project table is:

name = "project-name"
channels = ["conda-forge"]
platforms = ["linux-64"]


The name of the project.

name = "project-name"


This is a list that defines the channels used to fetch the packages from. If you want to use channels hosted on you only need to use the name of the channel directly.

channels = ["conda-forge", "robostack", "bioconda", "nvidia", "pytorch"]

Channels situated on the file system are also supported with absolute file paths:

channels = ["conda-forge", "file:///home/user/staged-recipes/build_artifacts"]

To access private or public channels on or Quetz use the url including the hostname:

channels = ["conda-forge", ""]


Defines the list of platforms that the project supports. Pixi solves the dependencies for all these platforms and puts them in the lock file (pixi.lock).

platforms = ["win-64", "linux-64", "osx-64", "osx-arm64"]

The available platforms are listed here: link

version (optional)#

The version of the project. This should be a valid version based on the conda Version Spec. See the version documentation, for an explanation of what is allowed in a Version Spec.

version = "1.2.3"

authors (optional)#

This is a list of authors of the project.

authors = ["John Doe <>", "Marie Curie <>"]

description (optional)#

This should contain a short description of the project.

description = "A simple description"

license (optional)#

The license as a valid SPDX string (e.g. MIT AND Apache-2.0)

license = "MIT"

license-file (optional)#

Relative path to the license file.

license-file = ""

readme (optional)#

Relative path to the README file.

readme = ""

homepage (optional)#

URL of the project homepage.

homepage = ""

repository (optional)#

URL of the project source repository.

repository = ""

documentation (optional)#

URL of the project documentation.

documentation = ""

conda-pypi-map (optional)#

Mapping of channel name or URL to location of mapping that can be URL/Path. Mapping should be structured in json format where conda_name: pypi_package_name. Example:

  "jupyter-ros": "my-name-from-mapping",
  "boltons": "boltons-pypi"

If conda-forge is not present in conda-pypi-map pixi will use mapping for it.

conda-pypi-map = { "conda-forge" = "", "" = "local/robostack_mapping.json"}

The tasks table#

Tasks are a way to automate certain custom commands in your project. For example, a lint or format step. Tasks in a pixi project are essentially cross-platform shell commands, with a unified syntax across platforms. For more in-depth information, check the Advanced tasks documentation. Pixi's tasks are run in a pixi environment using pixi run and are executed using the deno_task_shell.

simple = "echo This is a simple task"
cmd = { cmd="echo Same as a simple task but now more verbose"}
depending = { cmd="echo run after simple", depends-on="simple"}
alias = { depends-on=["depending"]}
download = { cmd="curl -o file.txt" , outputs=["file.txt"]}
build = { cmd="npm build", cwd="frontend", inputs=["frontend/package.json", "frontend/*.js"]}
run = { cmd="python $ARGUMENT", env={ ARGUMENT="value" }}
clean-env = { cmd = "python", clean-env = true} # Only on Unix!

You can modify this table using pixi task.


Specify different tasks for different platforms using the target table


If you want to hide a task from showing up with pixi task list or pixi info, you can prefix the name with _. For example, if you want to hide depending, you can rename it to _depending.

The system-requirements table#

The system requirements are used to define minimal system specifications used during dependency resolution. For example, we can define a unix system with a specific minimal libc version. This will be the minimal system specification for the project. System specifications are directly related to the virtual packages.

Currently, the specified defaults are the same as conda-lock's implementation:

default system requirements for linux
linux = "5.10"
libc = { family="glibc", version="2.17" }
default system requirements for windows
default system requirements for osx
macos = "10.15"
default system requirements for osx-arm64
macos = "11.0"

Only if a project requires a different set should you define them.

For example, when installing environments on old versions of linux. You may encounter the following error:

× The current system has a mismatching virtual package. The project requires '__linux' to be at least version '5.10' but the system has version '4.12.14'

This suggests that the system requirements for the project should be lowered. To fix this, add the following table to your configuration:

linux = "4.12.14"

Using Cuda in pixi#

If you want to use cuda in your project you need to add the following to your system-requirements table:

cuda = "11" # or any other version of cuda you want to use

This informs the solver that cuda is going to be available, so it can lock it into the lock file if needed.

The pypi-options table#

The pypi-options table is used to define options that are specific to PyPI registries. These options can be specified either at the root level, which will add it to the default options feature, or on feature level, which will create a union of these options when the features are included in the environment.

Alternative registries#

Currently the main reason to use this table is to define alternative registries. We support:

  • index-url: replaces the main index url. Only one index-url can be defined per environment.
  • extra-index-urls: adds an extra index url.
  • find-links: which can either be a path {path = './links'} or a url {url = ''}. This is similar to the --find-links option in pip.

An example:

index-url = ""
extra-index-urls = [""]
find-links = [{path = './links'}]

There are some examples in the pixi repository that make use of this feature. To read about existing authentication methods, please check the PyPI Authentication section.

Strict Index Priority

Unlike pip, because we make use of uv, we have a strict index priority. This means that the first index is used where a package can be found. The order is determined by the order in the toml file. Where the extra-index-urls are preferred over the index-url. Read more about this on the UV Readme

The dependencies table(s)#

This section defines what dependencies you would like to use for your project.

There are multiple dependencies tables. The default is [dependencies], which are dependencies that are shared across platforms.

Dependencies are defined using a VersionSpec. A VersionSpec combines a Version with an optional operator.

Some examples are:

# Use this exact package version
package0 = "1.2.3"
# Use 1.2.3 up to 1.3.0
package1 = "~=1.2.3"
# Use larger than 1.2 lower and equal to 1.4
package2 = ">1.2,<=1.4"
# Bigger or equal than 1.2.3 or lower not including 1.0.0
package3 = ">=1.2.3|<1.0.0"

Dependencies can also be defined as a mapping where it is using a matchspec:

package0 = { version = ">=1.2.3", channel="conda-forge" }
package1 = { version = ">=1.2.3", build="py34_0" }


The dependencies can be easily added using the pixi add command line. Running add for an existing dependency will replace it with the newest it can use.


To specify different dependencies for different platforms use the target table


Add any conda package dependency that you want to install into the environment. Don't forget to add the channel to the project table should you use anything different than conda-forge. Even if the dependency defines a channel that channel should be added to the project.channels list.

python = ">3.9,<=3.11"
rust = "1.72"
pytorch-cpu = { version = "~=1.1", channel = "pytorch" }


Details regarding the PyPI integration

We use uv, which is a new fast pip replacement written in Rust.

We integrate uv as a library, so we use the uv resolver, to which we pass the conda packages as 'locked'. This disallows uv from installing these dependencies itself, and ensures it uses the exact version of these packages in the resolution. This is unique amongst conda based package managers, which usually just call pip from a subprocess.

The uv resolution is included in the lock file directly.

Pixi directly supports depending on PyPI packages, the PyPA calls a distributed package a 'distribution'. There are Source and Binary distributions both of which are supported by pixi. These distributions are installed into the environment after the conda environment has been resolved and installed. PyPI packages are not indexed on but can be viewed on

Important considerations

  • Stability: PyPI packages might be less stable than their conda counterparts. Prefer using conda packages in the dependencies table where possible.
  • Compatibility limitation: Currently, pixi doesn't support private PyPI repositories

Version specification:#

These dependencies don't follow the conda matchspec specification. The version is a string specification of the version according to PEP404/PyPA. Additionally, a list of extra's can be included, which are essentially optional dependencies. Note that this version is distinct from the conda MatchSpec type. See the example below to see how this is used in practice:

# When using pypi-dependencies, python is needed to resolve pypi dependencies
# make sure to include this
python = ">=3.6"

fastapi = "*"  # This means any version (the wildcard `*` is a pixi addition, not part of the specification)
pre-commit = "~=3.5.0" # This is a single version specifier
# Using the toml map allows the user to add `extras`
pandas = { version = ">=1.0.0", extras = ["dataframe", "sql"]}

# git dependencies
# With ssh
flask = { git = "ssh://" }
# With https and a specific revision
requests = { git = "", rev = "0106aced5faa299e6ede89d1230bd6784f2c3660" }
# TODO: will support later -> branch = '' or tag = '' to specify a branch or tag

# You can also directly add a source dependency from a path, tip keep this relative to the root of the project.
minimal-project = { path = "./minimal-project", editable = true}

# You can also use a direct url, to either a `.tar.gz` or `.zip`, or a `.whl` file
click = { url = "" }

# You can also just the default git repo, it will checkout the default branch
pytest = { git = ""}

Full specification#

The full specification of a PyPI dependencies that pixi supports can be split into the following fields:


A list of extras to install with the package. e.g. ["dataframe", "sql"] The extras field works with all other version specifiers as it is an addition to the version specifier.

pandas = { version = ">=1.0.0", extras = ["dataframe", "sql"]}
pytest = { git = "URL", extras = ["dev"]}
black = { url = "URL", extras = ["cli"]}
minimal-project = { path = "./minimal-project", editable = true, extras = ["dev"]}

The version of the package to install. e.g. ">=1.0.0" or * which stands for any version, this is pixi specific. Version is our default field so using no inline table ({}) will default to this field.

py-rattler = "*"
ruff = "~=1.0.0"
pytest = {version = "*", extras = ["dev"]}

A git repository to install from. This support both https:// and ssh:// urls.

Use git in combination with rev or subdirectory:

  • rev: A specific revision to install. e.g. rev = "0106aced5faa299e6ede89d1230bd6784f2c3660
  • subdirectory: A subdirectory to install from. subdirectory = "src" or subdirectory = "src/packagex"
# Note don't forget the `ssh://` or `https://` prefix!
pytest = { git = ""}
requests = { git = "", rev = "0106aced5faa299e6ede89d1230bd6784f2c3660" }
py-rattler = { git = "ssh://", subdirectory = "py-rattler" }

A local path to install from. e.g. path = "./path/to/package" We would advise to keep your path projects in the project, and to use a relative path.

Set editable to true to install in editable mode, this is highly recommended as it is hard to reinstall if you're not using editable mode. e.g. editable = true

minimal-project = { path = "./minimal-project", editable = true}

A URL to install a wheel or sdist directly from an url.

pandas = {url = ""}
Did you know you can use: add --pypi?

Use the --pypi flag with the add command to quickly add PyPI packages from the CLI. E.g pixi add --pypi flask

This does not support all the features of the pypi-dependencies table yet.

Source dependencies (sdist)#

The Source Distribution Format is a source based format (sdist for short), that a package can include alongside the binary wheel format. Because these distributions need to be built, the need a python executable to do this. This is why python needs to be present in a conda environment. Sdists usually depend on system packages to be built, especially when compiling C/C++ based python bindings. Think for example of Python SDL2 bindings depending on the C library: SDL2. To help built these dependencies we activate the conda environment that includes these pypi dependencies before resolving. This way when a source distribution depends on gcc for example, it's used from the conda environment instead of the system.


This table contains dependencies that are needed to build your project but which should not be included when your project is installed as part of another project. In other words, these dependencies are available during the build but are no longer available when your project is installed. Dependencies listed in this table are installed for the architecture of the target machine.

python = "~=3.10.3"

Typical examples of host dependencies are:

  • Base interpreters: a Python package would list python here and an R package would list mro-base or r-base.
  • Libraries your project links against during compilation like openssl, rapidjson, or xtensor.


This table contains dependencies that are needed to build the project. Different from dependencies and host-dependencies these packages are installed for the architecture of the build machine. This enables cross-compiling from one machine architecture to another.

cmake = "~=3.24"

Typical examples of build dependencies are:

  • Compilers are invoked on the build machine, but they generate code for the target machine. If the project is cross-compiled, the architecture of the build and target machine might differ.
  • cmake is invoked on the build machine to generate additional code- or project-files which are then include in the compilation process.


The build target refers to the machine that will execute the build. Programs and libraries installed by these dependencies will be executed on the build machine.

For example, if you compile on a MacBook with an Apple Silicon chip but target Linux x86_64 then your build platform is osx-arm64 and your host platform is linux-64.

The activation table#

The activation table is used for specialized activation operations that need to be run when the environment is activated.

There are two types of activation operations a user can modify in the manifest:

  • scripts: A list of scripts that are run when the environment is activated.
  • env: A mapping of environment variables that are set when the environment is activated.

These activation operations will be run before the pixi run and pixi shell commands.


The activation operations are run by the system shell interpreter as they run before an environment is available. This means that it runs as cmd.exe on windows and bash on linux and osx (Unix). Only .sh, .bash and .bat files are supported.

And the environment variables are set in the shell that is running the activation script, thus take note when using e.g. $ or %.

If you have scripts or env variable per platform use the target table.

scripts = [""]
env = { ENV_VAR = "value" }

# To support windows platforms as well add the following
scripts = ["env_setup.bat"]

ENV_VAR = "linux-value"

The target table#

The target table is a table that allows for platform specific configuration. Allowing you to make different sets of tasks or dependencies per platform.

The target table is currently implemented for the following sub-tables:

The target table is defined using [target.PLATFORM.SUB-TABLE]. E.g [target.linux-64.dependencies]

The platform can be any of:

  • win, osx, linux or unix (unix matches linux and osx)
  • or any of the (more) specific target platforms, e.g. linux-64, osx-arm64

The sub-table can be any of the specified above.

To make it a bit more clear, let's look at an example below. Currently, pixi combines the top level tables like dependencies with the target-specific ones into a single set. Which, in the case of dependencies, can both add or overwrite dependencies. In the example below, we have cmake being used for all targets but on osx-64 or osx-arm64 a different version of python will be selected.

cmake = "3.26.4"
python = "3.10"

python = "3.11"

Here are some more examples:

scripts = ["setup.bat"]

msmpi = "~=10.1.1"

vs2022_win-64 = "19.36.32532"

tmp = "echo $TEMP"

clang = ">=16.0.6"

The feature and environments tables#

The feature table allows you to define features that can be used to create different [environments]. The [environments] table allows you to define different environments. The design is explained in the this design document.

Simplest example
pytest = "*"

test = ["test"]

This will create an environment called test that has pytest installed.

The feature table#

The feature table allows you to define the following fields per feature.

  • dependencies: Same as the dependencies.
  • pypi-dependencies: Same as the pypi-dependencies.
  • system-requirements: Same as the system-requirements.
  • activation: Same as the activation.
  • platforms: Same as the platforms. Unless overridden, the platforms of the feature will be those defined at project level.
  • channels: Same as the channels. Unless overridden, the channels of the feature will be those defined at project level.
  • target: Same as the target.
  • tasks: Same as the tasks.

These tables are all also available without the feature prefix. When those are used we call them the default feature. This is a protected name you can not use for your own feature.

Full feature table specification
activation = {scripts = [""]}
# Results in:  ["nvidia", "conda-forge"] when the default is `conda-forge`
channels = ["nvidia"]
dependencies = {cuda = "x.y.z", cudnn = "12.0"}
pypi-dependencies = {torch = "==1.9.0"}
platforms = ["linux-64", "osx-arm64"]
system-requirements = {cuda = "12"}
tasks = { warmup = "python" }
target.osx-arm64 = {dependencies = {mlx = "x.y.z"}}
Full feature table but written as separate tables
scripts = [""]

cuda = "x.y.z"
cudnn = "12.0"

torch = "==1.9.0"

cuda = "12"

warmup = "python"

mlx = "x.y.z"

# Channels and Platforms are not available as separate tables as they are implemented as lists
channels = ["nvidia"]
platforms = ["linux-64", "osx-arm64"]

The environments table#

The [environments] table allows you to define environments that are created using the features defined in the [feature] tables.

The environments table is defined using the following fields:

  • features: The features that are included in the environment. Unless no-default-feature is set to true, the default feature is implicitly included in the environment.
  • solve-group: The solve group is used to group environments together at the solve stage. This is useful for environments that need to have the same dependencies but might extend them with additional dependencies. For instance when testing a production environment with additional test dependencies. These dependencies will then be the same version in all environments that have the same solve group. But the different environments contain different subsets of the solve-groups dependencies set.
  • no-default-feature: Whether to include the default feature in that environment. The default is false, to include the default feature.

Full environments table specification
test = {features = ["test"], solve-group = "test"}
prod = {features = ["prod"], solve-group = "test"}
lint = {features = ["lint"], no-default-feature = true}
As shown in the example above, in the simplest of cases, it is possible to define an environment only by listing its features:

Simplest example
test = ["test"]

is equivalent to

Simplest example expanded
test = {features = ["test"]}

When an environment comprises several features (including the default feature): - The activation and tasks of the environment are the union of the activation and tasks of all its features. - The dependencies and pypi-dependencies of the environment are the union of the dependencies and pypi-dependencies of all its features. This means that if several features define a requirement for the same package, both requirements will be combined. Beware of conflicting requirements across features added to the same environment. - The system-requirements of the environment is the union of the system-requirements of all its features. If multiple features specify a requirement for the same system package, the highest version is chosen. - The channels of the environment is the union of the channels of all its features. Channel priorities can be specified in each feature, to ensure channels are considered in the right order in the environment. - The platforms of the environment is the intersection of the platforms of all its features. Be aware that the platforms supported by a feature (including the default feature) will be considered as the platforms defined at project level (unless overridden in the feature). This means that it is usually a good idea to set the project platforms to all platforms it can support across its environments.

Global configuration#

The global configuration options are documented in the global configuration section.