Skip to main content


The @conda decorator specifies what libraries should be made available for a step.

The libraries are installed from Conda repositories. For more information, see Managing External Libraries.



from metaflow import conda

Specifies the Conda environment for the step.

Information in this decorator will augment any attributes set in the @conda_base flow-level decorator. Hence, you can use @conda_base to set packages required by all steps and use @conda to specify step-specific overrides.


packages: Dict[str, str], default {}

Packages to use for this step. The key is the name of the package and the value is the version to use.

libraries: Dict[str, str], default {}

Supported for backward compatibility. When used with packages, packages will take precedence.

python: str, optional, default None

Version of Python to use, e.g. '3.7.4'. A default value of None implies that the version used will correspond to the version of the Python interpreter used to start the run.

disabled: bool, default False

If set to True, disables @conda.