This document explains how to access data hosted by NeuroData. Data are publicly available. NeuroData utilizes the BOSS, co-developed with JHU-APL, as a data store. We also selectively host small datasets on S3 for fast data visualization in neuroglancer. For that, see our guide on the precomputed format.

Boss hierarchy

  1. Collection: typically either the name of the lab PI or the person who collected the data
  2. Experiment: within this level, data share a common voxel extent and voxel size. Often synonymous with actual experiments.
  3. Channel: data reside at this level and are either uint8/uint16 image data or uint64 annotations

Further information on BOSS organization can be found in the docs.

Methods for accessing data

Account needed

All data access currently requires account registration. Create an account at

There are four methods for getting data from the BOSS:


A frontend for navigating BOSS projects. It provides TIFF cutouts (limited to < 1GB) to data and links to data visualization (through neuroglancer). Larger cutouts should be performed programmatically using either ndex or intern detailed below.

  • Log in at
  • Navigate the available projects by clicking the links to collections and experiments. Clicking on any experiment takes you to the cutout page.
  • The cutout page provides TIFF downloads and viz links for channels within each experiment.


To visualize the data before downloading you can use neuroglancer. Please see our guide.


A Python command-line tool for downloading volumes of data from the BOSS as TIFF stacks. This library can also be loaded from within Python.

Installation and usage instructions on its GitHub page


A Python package developed by JHU-APL for interactive and programmatic usage of the BOSS.

Installation and usage instructions available on its GitHub page