Scipp requires Python 3.7 or above.


The easiest way to install scipp is using conda. Packages from Anaconda Cloud are available for Linux, macOS, and Windows. It is recommended to create an environment rather than installing individual packages.

With the provided environment file#

  1. Download scipp.yml for the stable release version of scipp. Use scipp-latest.yml for the latest development version.

  2. In a terminal run:

    conda activate
    conda env create -f scipp.yml
    conda activate scipp
    jupyter lab

    The conda activate ensures that you are in your base environment. This will take a few minutes. Above, replace scipp.yml with the path to the download location you used to download the environment. Open the link printed by Jupyter in a browser if it does not open automatically.

If you are new to scipp, continue reading with Quick Start and Data Structures.

If you have previously installed scipp with conda we nevertheless recommend creating a fresh environment rather than trying to conda update. You may want to remove your old environment first, e.g.,

conda activate
conda env remove -n scipp

and then proceed as per instructions above. The conda activate ensures that you are in your base environment.

Without the provided environment file#

To create a new conda environment with scipp:

conda create -n env_with_scipp -c conda-forge -c scipp scipp

To add scipp to an existing conda environment:

conda install -c conda-forge -c scipp scipp

For a more up-to-date version, the scipp/label/dev channel can be used instead:

conda install -c conda-forge -c scipp/label/dev scipp


Installing scipp on Windows requires Microsoft Visual Studio 2019 C++ Runtime (and versions above) installed. Visit for the up to date version of the library.

After installation the module scipp can be imported in Python. Note that only the bare essential dependencies are installed. If you wish to use plotting functionality you will also need to install matplotlib and ipywidgets.

To update or remove scipp use conda update and conda remove.


Scipp is available from PyPI via pip:

pip install scipp

By default, this is only a minimal install without optional dependencies. To install components for additional submodules, use the all extras of the module:

pip install scipp[all]

This will install everything that is needed, with the exception of plotting libraries for interactive use in Jupyter. These are avaiable in the interactive extra. We recommend combining this with all:

pip install scipp[all,interactive]

From source#

See developer getting started.