Zarr Viewer
Open Zarr v2 and v3 stores in your browser, drop the store folder or a .zip. Browse arrays and groups, slice N-dimensional data, and render heatmaps, reading only the chunks you view. Nothing is uploaded.
Free, no sign-up, and fully client-side. Your file never leaves your browser. Opens .zarr (folder or .zip).
Open the viewer and drop a file to get started.
What is Zarr?
Zarr is a format for chunked, compressed, N-dimensional arrays. It is popular in geospatial and climate science, bio-imaging and microscopy, and machine learning, commonly paired with xarray and the Pangeo stack. A Zarr store is a directory (or a zip of one) of chunk files plus small JSON metadata files (.zarray/.zgroup for v2, zarr.json for v3).
Rather than spinning up xarray just to inspect a store, drop the folder here and browse it immediately. ViewKit fetches only the chunks a given slice needs, decompresses them in your browser, and never uploads anything.
What you can do with the Zarr viewer
Zarr v2 and v3
Reads both the v2 layout (.zarray/.zgroup) and the v3 layout (zarr.json) transparently.
Folder or zip
Open a Zarr directory with the browser’s folder picker, or drop a .zip of the store.
Chunk-aware reads
Only the chunks intersecting your slice are fetched and decompressed; a small cache and neighbor prefetch keep scrubbing smooth.
Heatmaps & tables
Visualize any 2-D slice of an N-D array as a heatmap or a table using the dimension slicer.
How to open a Zarr file
- 1
Use “open a Zarr directory” above and pick the store folder, or drop a .zip of the store. It is read in your browser, nothing uploads.
- 2
Browse the groups and arrays; click an array to see its shape, dtype and chunking.
- 3
Slice N-dimensional data with the dimension mapper and switch between a table and a heatmap.
Doing the same thing in code
The equivalent in Python, using zarr:
import zarr
store = zarr.open("data.zarr", mode="r")
print(store.tree()) # groups and arrays
arr = store["temperature"]
print(arr.shape, arr.dtype, arr.chunks, arr.compressor)
print(arr[0, :5, :5]) # only the intersecting chunks are readThe installed version has to match the store. zarr-python 2 cannot read a v3 store, and a codec that is not installed fails at read time rather than on open. This page reads both the v2 and v3 layouts.
Zarr viewer: frequently asked questions
Does it support Zarr v3?
Yes, both Zarr v2 and Zarr v3 stores are supported.
How do I tell whether a store is v2 or v3?
Check the metadata files at the top of the store. A v2 store has .zgroup and .zarray files; a v3 store has a single zarr.json per node. zarr-python 2 cannot open a v3 store, which is a common cause of a store failing to open in an older environment.
Do I need to zip the store first?
No. Use the directory option and select the store folder, and the whole tree is read in place. A .zip of the store also works.
How do I open a Zarr directory in the browser?
Use the “open a Zarr directory” option and select the store folder, or drop a .zip archive of the store onto the page.
Which compression codecs work?
Common Zarr codecs such as Blosc, Zstd, Gzip and LZ4 are supported.
Is my data uploaded?
No. Chunks are fetched and decoded in your browser; nothing about your store is sent to a server.