Skip to content

Downloading Data and Dandisets

You can download the content of a Dandiset using the DANDI Web application (such a specific file) or entire Dandisets using the DANDI Python CLI.

Using the DANDI Web Application

Once you have the Dandiset you are interested in (see more in the Dandiset View section), you can download the content of the Dandiset. On the landing page of each Dandiset, you can find Download button on the right-hand panel. After clicking the button, you will see the specific command you can use with DANDI Python CLI (as well as the information on how to download the CLI).


Download specific files

The right-side panel of the Dandiset landing page allows you also to access the list of folders and files.


Each file in the Dandiset has a download icon next to it, clicking the icon will start the download process.

Using the Python CLI Client

The DANDI Python client gives you more options, such as downloading entire Dandisets.

Before You Begin: You need to have Python 3.8+ and install the DANDI Python Client using pip install dandi. If you have an issue using the Python CLI, see the Dandi Debugging section.

Download a Dandiset

To download an entire Dandiset, you can use the same command as suggested by DANDI web application, e.g.:

dandi download DANDI:000023

Download data for a specific subject from a Dandiset

You can download data for specific subjects. Names of the subjects can be found on DANDI web application or by running a command with the DANDI CLI: dandi ls -r DANDI:000023. Once you have the subject ID, you can download the data, e.g.:

dandi download

You should replace _draft_ with a specific version you are interested in (e.g. 0.210914.1900 in the case of this Dandiset).

You can also use the link from DANDI web application, e.g.:

dandi download

Download a specific file from a Dandiset

You can download a specific file from a Dandiset when the link for the specific file can be found on the DANDI web application, e.g.:

dandi download

Hint: dandi download supports a number of resource identifiers to point to a Dandiset, folder, or file. Providing an incorrect URL (e.g. dandi download wrongurl) will provide a list of supported identifiers.

Using DataLad

All dandisets are regularly mirrored to DataLad datasets which are made available at the GitHub organization Where present, individual Zarr files are included as subdatasets (git submodules) hosted in the GitHub organization

The Git revision histories of each dataset reflect the Dandiset's draft state as of each execution of the mirroring job. Published Dandiset versions are tagged with Git tags.

With DataLad, you can: - clone an entire dataset, - use a specific version of it, - explore history of modifications, - download content of files of interest, - locally discard the content of no-longer-needed files, - use the dataset in a reproducible manner, - include it as a subdataset in your own DataLad dataset, - use to FUSE-mount individual locally-cloned dandisets so that their files' contents are transparently streamed to your DANDI/DataLad-unaware tools, - etc.

Learn more about DataLad from its handbook at

Developers' note: DataLad datasets are created using the dandi/backups2datalad tool which is also available for use by the community to similarly maintain mirrors of independent DANDI deployments as DataLad datasets.

Using WebDAV

DANDI provides a WebDAV service at for accessing the data in the DANDI archive. You can use any WebDAV client or even a web browser to access the data - any dandiset, any version, any file or collection of files. You can use any web download tool to download the data from the DANDI archive, e.g.

wget -r -np -nH --cut-dirs=3

for a download of a specific release 0.210831.2033 of the 000027 dandiset.

Note: The WebDAV service does not directly serve any file contents; it instead relies on redirects to AWS S3 storage where the contents are stored. You might need to configure your WebDAV client to follow redirects; e.g., for the davfs2 WebDAV client, set follow_redirect to 1 in /etc/davfs2/davfs2.conf.

Developers' note: The WebDAV service's code is available at and can also be used for independent DANDI deployments.