Datasets As First-class Resources

Position

Datasets are first-class resources in Prox OS. They are not just pages inside an app and they are not only connector cache rows. A Dataset can mount into a Studio, appear in an App, inform Alma, and later become a shareable or owner-scoped resource.

Current Hub

/datasets is the v0 Dataset Hub. It uses local mock data to show:

• dataset cards;
• visibility badges: public, private, shared, connector-backed;
• schema previews;
• activity and permission placeholders;
• how a dataset mounts into Studio Engines.

Mock examples include Moments, AI Questions, Visitor Logs, Contacts, Calendar, and Subscriptions.

Mounting Direction

A Studio Blueprint can reference datasets by id, permission, mount type, and presentation. A future runtime should resolve:

• who owns the dataset;
• whether the current user can read, edit, share, or export it;
• which connector or source produced it;
• which Studio, App, or AI context can use it;
• how it appears in Desktop Runtime, Atlas, Grid, Ops, local-only development engines, or AI Rail surfaces.

No real connector OAuth or backend storage is required for the v0 hub.