Data

The Data app is one tree for every file you can reach. Your cloud storage, anything you have connected, and this computer all sit in it side by side, with the same right-click menu wherever the bytes actually live.

The tree

Three kinds of root, in this order:

• Relay Drive — your cloud storage. Files here are safe to share, run pipelines on, and open from another machine.
• Connected sources — Google Drive, a lab bucket, DANDI, and so on. Each appears as its own root once connected. Nothing is copied until you open a file.
• Local Disk — this computer. In the browser, files you drop here are kept in the browser’s own storage; in the desktop app it is your real disk.

Teams add a Team Drive root that every member sees.

Getting files in

Drag files or folders onto the Data window, or straight onto its dock icon. They land in Local Disk and are usable at once. Large files upload to your Drive in the background while you keep working.

The Add button (the + in the toolbar) offers Upload files, Upload folder and Connect a source.

The Upload Queue panel shows what is in flight. Uploads retry on their own after a network hiccup, and you can pause, resume, or retry a failed batch from there. Files are sent in chunks, so a big one does not start over after a retry.

Note

Two folders in your Drive fill themselves in as you work. Notebooks/ holds every notebook you create, with version history. Studio/ holds the presentations, papers and posters you build. Pipeline results go to pipeline-results/ unless you choose another folder on the pipeline’s Export node.

Opening and previewing

Selecting a file previews it: real dimensions, channels and size, read from the file rather than guessed from its extension. Double-click, or press Open, to open it. Notebooks open in Notebooks, Studio documents in Studio, everything else in its own Viewer window. What each file type opens as is on the Viewers page.

Select several files, or a folder, and right-click for Open synchronized viewer to put recordings from one experiment on a shared clock. See Time synchronization.

The right-click menu

On a file: Download, Copy Name, Copy Path, Details, Rename, Share, Export as, and Delete.

On a folder: Arrange as plate (for a folder of well images), Export All as OME-Zarr, Auto-organize (Ray proposes a folder structure for what is inside, which you approve before anything moves), Share Folder, Create Folder, Rename, Delete Folder, Details and Refresh. A folder of videos in Relay Drive or Local Disk also offers Create pose training set…; see Pose labels.

A file with annotations waiting for a decision carries an N to review badge in the tree. See Needs review.

Warning

Delete is permanent. There is no recycle bin in Relay Drive.

Connecting a source

Add → Connect a source opens the Integrations picker. Pick a source, sign in or paste a key, and it appears in the tree. The picker is grouped by what you have:

Group	Sources	What you provide
Cloud storage	Google Drive, Dropbox, Box, OneDrive, SharePoint	Sign in
	Amazon S3, Wasabi, Backblaze B2	Access keys
	Google Cloud Storage, Azure Blob Storage	A key or token
Published data	Zenodo, Figshare, OSF, OpenNeuro	API token
	Dryad, AWS Open Data Registry	Nothing
Genomics	NCBI SRA / GEO	API token
Neuroscience	DANDI Archive	API token
Imaging	OMERO	Host, username and password
Code & models	GitHub, Hugging Face Datasets	Sign in or token
HPC transfer	Globus	Globus sign-in

You can also paste a direct URL to a file, a Zarr store or a Neuroglancer dataset with Paste a Dataset URL.

Tip

Public archives like DANDI, IDR and the Allen Institute are also browsable without connecting anything, through Commons.

Your own bucket

If your lab’s data already lives in a bucket, add it instead of uploading. Connect a source → Amazon S3 Bucket (or the S3-compatible option) walks through three steps. Google Cloud Storage and Azure Blob Storage appear in the list but are not supported as your own bucket yet.

1. Choose the provider

   AWS S3, or S3-compatible (on-prem) for Ceph, MinIO, Synology and similar gateways.

2. Enter the bucket details

   A name for the connection, the bucket name, the region, and for S3-compatible storage the endpoint URL. Public buckets need no credentials.

3. Test the connection

   Relay lists the bucket to confirm it can read it, and tells you if the bucket’s CORS settings need a change.

Files in your bucket open, stream and share like any other. Credentials are encrypted and never shown again; remove the connection from the tree to revoke them.

How much space you have

Settings → Plan & usage → Usage shows how much of your Drive is used and what is left. Uploads that would go over your plan’s storage are refused before they start, so nothing is half-uploaded.