Using Zarr Datasets with webKnossos

Tom Herold
Published in
3 min readOct 17, 2022

We see growing support for the Zarr file format and OME-Zarr NGFF standards in the life sciences community and are working hard to make webKnossos work seamlessly with those.

In addition to exposing any datasets and annotations in webKnossos through the Zarr API to third-party tools (see my blog post), we also want to make it easy for people to load any Zarr-format-based datasets into webKnossos. We support reading Zarr dataset from the local file system, cloud providers such as S3 and Google Cloud, as well as through any web server over HTTP.

Zarr-based data sources will be supported in addition to the already existing support for WKW, Neuroglancer Precomputed, and BossDB sources.

Why should I work with my Zarr datasets in webKnossos?

We believe webKnossos has many useful and easy-to-use tools and functionalities for the life science community when working with Zarr datasets:

  • Like the Zarr file format, webKnossos is engineered to explore and work with large-scale volumetric datasets (from megabytes up to petabytes). It efficiently streams Zarr datasets and makes viewing your data super speedy.
  • webKnossos can serve as a central hub for your datasets. Once imported, each dataset can be tagged, searched, and shared with your colleagues. Unlike other software tools, there is no need to permanently remember and share long, cryptic URLs.
  • In addition to a great dataset viewing experience, webKnossos also comes with many tools for volume and skeleton annotation out-of-the-box. You can get started annotating right away without having to switch between viewing and editing tools.
  • webKnossos is built around teams, sharing, and authentication. If you want to work in a collaborative environment, e.g. colleagues in your lab, webKnossos already has all the essentials ready for that.
  • Any dataset and annotation available in webKnossos can also be streamed to other OME-Zarr compatible software tools. There is no lock-in.

How do I open my Zarr dataset in webKnossos?

webKnossos makes it easy to load Zarr datasets from remote sources, i.e. cloud storage or a public web server (HTTP).

You can add a new dataset from your dashboard:

  • Click the “Add Dataset” button under the “Datasets” tab
  • Switch to the “Add Remote Zarr Dataset” tab
  • Enter the URL to a valid Zarr source and click “Add Layer”
    (Provide authentication credentials in case of private cloud storage access. Not required for public datasets)
  • webKnossos will create a new dataset and add the above source a new layer to it. Further, wK will auto-detect the layer name, data types, and related metadata from the remote source.
  • Add other Zarr sources as additional layers as needed, e.g., one layer for the “color” data from the microscope, one layer containing the segmentation, etc.
  • Click “Import” to finish the process. The remote dataset should now be listed with the rest of your datasets in the dashboard and is ready for the typical webKnossos operations.
The webKnoss user interface for importing Zarr datasets. A dataset can be composed from several remote Zarr sources.

Note, webKnossos will only save and manage related metadata for any imported Zarr dataset and not make a physical copy of the data itself. Any access to the imported dataset will stream the data into webKnossos from its remote source.

We believe that establishing standardized file formats, such as OME-Zarr, for working with volumetric image data is moving the whole life sciences community forward. We are proud that webKnossos now natively supports the new OME-Zarr as one of the first analysis tool for scientist.
Get started for free on

Please contact us for feedback or if you run into any issues with your Zarr datasets.