Software we have built outside the NWB toolchain, for working with cloud-hosted data and for applying and tracking language models.
General-purpose libraries for reading and writing large scientific arrays in the cloud, which grew out of our NWB work but are not specific to it.
An HDF5 virtual file driver that streams remote files over HTTP, fetching only the bytes a read needs with adaptive caching. It is a C++ counterpart to the Python remfile package, works with any HDF5 build, and opens remote NWB files several times faster than the ROS3 driver. It is an optional dependency of AqNWB, the C++ API for NWB.
A pure-MATLAB implementation of the Zarr v3 specification, for chunked, compressed N-dimensional arrays stored locally or in the cloud. It supports every Zarr data type, zstd, blosc, and gzip compression, sharding, and read-only access over HTTP, and its continuous integration checks on every commit that MATLAB and zarr-python can read each other's files. hdmf-zarr-matlab and matzarr build on it to read NWB Zarr files and cloud-hosted .mat files from MATLAB.
Tools that apply language models to finding neuroscience data, and that track the cost and capability of the models themselves.
A browsable map of the DANDI Archive. Every public Dandiset is placed by the meaning of its metadata, using sentence embeddings of its title, description, and experimental details, so datasets that describe similar science sit near one another even when they use different vocabulary. Topic regions are discovered by clustering, points can be colored by topic or species, and the map rebuilds itself from the DANDI API every night.
An agentic science project, in which AI agents carry out scientific work end to end. Encore Science is a public registry of attempts to reproduce the key findings of published papers from their open data, with agents doing the work. An agent reads each paper and writes its claims down before running any analysis, then tries to reproduce them from the dataset the paper deposited, starting with the DANDI Archive. Independent agents check every attempt, and each claim gets a recorded outcome whether or not it reproduced, along with an executed notebook and any problems found in the data itself. A failure to reproduce may be a failure of the agent and is not a judgement of the paper. The same checks can be run on a lab's own data with the encore command.
A dashboard that tracks how cheaply each level of language-model capability can be bought. It plots the Pareto frontier of the Artificial Analysis Intelligence Index against measured cost per task, shows how that frontier has moved over time, and lists the cheapest model that reaches each capability tier. The data refreshes every week.
