Build a connected workspace.
Start with the sample, or build your own model from nothing. The records, relationships, queries, and revisions are all working parts of the browser application.
Open the workspaceExplore the sample
The initial Northstar workspace contains 34 records, 31 relationships, and four synthetic source snapshots. It connects customer accounts and contacts with products, subscription plans, services, and incidents.
Open Query and choose Find exposed enterprise accounts. The initial sample returns Aster Labs, Copperline, and Fieldwork. Inspect a result to see its properties, relationships, and latest recorded source.
Open Data sources and apply Resolve the payment incident. Run the same query again. It now returns Aster Labs and Fieldwork, whose Analytics subscriptions still depend on degraded Reporting. Copperline no longer matches.
Choose Revision 4 in the revision selector and rerun the query. The previous three-account result returns. This is a query over the earlier snapshot, not a saved answer.
Start from empty
Select Build from empty, name the workspace, and confirm. This replaces the current local workspace. Export a backup first to keep existing work.
You can define your own model immediately, or import the sample sources in this order: Customer directory, Application catalog, Subscription ledger, then Service observations. Relationship imports need their participant records to exist, so an out-of-order import fails without partially writing the dataset.
Import your own records
Choose Data sources → Import your data. Paste an array of flat JSON records or select a CSV file. Preview the import, choose a record type and identifier column, then validate and import. The demo derives field types for a new model and checks compatibility with an existing type.
id,name,status,owner
asset:gateway,Gateway,active,Platform
asset:queue,Job queue,active,Applications
asset:archive,Archive,maintenance,DataTo connect the imported records, define a relationship type in Data model, name each participant role, and select its record type. Then use Connect records to choose the actual participants.
Keep your work
Workspace data is saved to this browser on a best-effort basis. Use the workspace menu to export the complete JSON backup, including the model, records, relationships, sources, and revision history. Importing a backup validates it before replacing the workspace.
Saved query shortcuts are separate browser preferences. Copy the JSON of any query you want to retain independently. Clearing browser data or changing the hosting origin can remove or isolate local state.
When another tab changes the saved workspace, editing is blocked rather than silently overwriting that tab’s data. Export the current session if needed, then reload the saved workspace.
Keep a local copy
Open the interactive demo in this browser, or download the sample asset CSV for the import walkthrough. The workspace is the same reviewed local engine, not a hosted client.
What this demo is
The workspace uses a local reference engine. It is not connected to a hosted tenant, live data connector, identity provider, payment system, or language model. The query format documented here belongs to this demo, not an official production API.
| Local demo limit | Value |
|---|---|
| Records | 1,500 |
| Relationships | 4,000 |
| Source snapshots | 300 |
| Retained revisions | 80, including the initial revision |
| Flat import | 500 records and 1 MB of input |
| Full graph view | Up to 120 visible records |
| Query results | Up to 500 root records, with bounded traversal |
These are demonstration limits, not SolarflareDB service limits. Browser storage may fill before a model limit is reached, particularly because the demo retains complete snapshots.