# apache-airflow-providers-pixi > Apache Airflow 3 provider to run Python callables in Pixi environments. Apache Airflow 3 provider that runs Python callables, Bash commands, manifest tasks, branch choices and sensor checks inside [Pixi](https://pixi.sh) environments, on the worker or in a Kubernetes pod. Install with `pip install apache-airflow-providers-pixi` (add the `[cncf.kubernetes]` extra for the pod operator); it needs Apache Airflow 3.1.2 or newer, and the workers need pixi 0.81.0 or newer, which the provider never installs itself. An environment is a Pixi project (`pixi_project_path`), a manifest file (`pixi_toml_path`), or an inline manifest (`dependencies`, `pypi_dependencies`). Functions are shipped as source, so they must be self-contained: imports inside the function, no variables from enclosing functions. ## Basics - [Home](): Installation, requirements and a quick start - [Changelog](): Changes per release - [Security](): Security considerations: installing pixi, run-time dependency resolution, pickle, caches, credentials ## Guides - [Pixi Connection](): The pixi connection type: credentials for private conda channels and PyPI indexes - [Pixi Operator](): PixiOperator: choosing the environment, the lock file, platforms, inline manifests, arguments, the Airflow context, return values, errors, environment variables from Variables and Connections, logs, timeouts, cache directories - [Pixi Task Decorator](): @task.pixi and how to write a function that is shipped to the environment - [Pixi Bash Operator](): PixiBashOperator and @task.pixi_bash: Bash commands with pixi run, and their environment variables - [Pixi Task Operator](): PixiTaskOperator: run a [tasks] entry of a pixi manifest with pixi run [args] - [Pixi Kubernetes Pod Operator](): PixiKubernetesPodOperator and @task.pixi_kubernetes: Pixi environments in Kubernetes pods, deferrable mode, size limits, errors - [Pixi Branch and Short Circuit Operators](): PixiBranchOperator, PixiShortCircuitOperator, @task.pixi_branch and @task.pixi_short_circuit: choose or skip downstream tasks with a callable in a Pixi environment - [Pixi Sensor](): PixiSensor and @task.pixi_sensor: sensor checks inside a Pixi environment - [Deployment](): Deploying: pixi in the worker image, shared caches, env_cache_path, projects in DAG bundles, private channels, pods - [Building on the Pixi Operator](): Building other providers' operators and decorators on the Pixi base classes ## Development - [System Tests](): Example DAGs and how to run the unit, integration and system tests - [Installing from sources](): Installing from sources, the local Airflow, building the docs, the CI Airflow matrix and the release process ## Optional - [Python API](): Python API reference of every operator, sensor, decorator, hook, exception and base class