Data and storage
Bring inputs into a Studio project from the Data Sources menu, and browse stored files in the storage browser.
Studio has no file tree beside the canvas. To bring data into a project, you use the Data Sources menu in the bottom toolbar. Each entry either stores the files or hands them straight to a tool window. To browse what is already stored, you use the storage browser at /data.
The Data Sources menu
| Entry | What it does |
|---|---|
| Upload & Analyze | Uploads files to storage. You can stop there, or send the files on to an AI Data Analysis tool window |
| Example Workflows | Loads example files into an AI Data Analysis window |
| Database Connections | Submits a search to an external database, then opens a Discovery Search window on that job |
| Recent Data | Picks files you used recently and passes them to a tool window |
| Import from External | Imports files from an external source into an AI Data Analysis window |
| Format Converters | Converts files between formats |
The left sidebar's Data Sources panel has Upload Files, Databases, Examples, Recent and Convert. Those buttons do nothing in the current build, so use the bottom toolbar's Data Sources menu instead.
Upload limits
The upload dialog takes up to 20 files per upload and 100 MB per file. It accepts these types:
| Kind | Extensions |
|---|---|
| Structures and molecules | .pdb, .sdf, .mol, .mol2, .xyz |
| Sequences and text | .fasta, .fa, .seq, .smi, .smiles, .txt |
| Tables | .csv, .tsv, .xls, .xlsx |
| Structured data | .json, .xml |
| Archives | .zip, .tar, .gz |
Each file goes to storage through a presigned URL, and then the platform registers it.
The storage browser
/data is a full-page browser for your stored files, outside Studio and without its experimental warning. /data/project/<project-id>, /data/campaign/<campaign-id> and /data/chat/<chat-id> open the same browser scoped to that entity.
- Scope. A member of one organization sees their own files. An organization admin browses the whole organization's storage. A member of several organizations picks one from the Organization selector.
- Folders. Folders follow the platform's storage layout. Known folders get readable names such as Files, Outputs, Tool Outputs, Campaigns, Chats, Exports and Uploads. In the breadcrumbs, campaign and chat folders appear as a shortened ID.
- Navigation. Breadcrumbs lead back up the tree, Filter files… narrows the current folder by name, and the refresh button reloads it.
- File actions. Select a file to see its details and these actions:
- Download
- Copy S3 URI
- Copy presigned link
- Open in Viewer, for file types the platform can display in place
The right sidebar in a Studio project has a Files tab. In the current build it shows "Files panel requires authentication" even when you are signed in, because the project page does not give it an organization. Use /data instead.
Getting results out
Tool results and job outputs are stored with the job. You download them from the Job Details window or the Jobs Dashboard. To export the whole project as a ZIP, CSV or JSON file, use File → Export Project Data. See Job management.
Was this page helpful?
Job management
Submit a job from a tool window or a parameter sweep, follow it in the Jobs Dashboard, and act on one job or many.
How the orchestrator picks tools
You describe a goal in plain language and the orchestrator maps it to tools and chains them. This page covers what happens in between and how to steer it.