Settings
Settings is an overlay panel (like everything else in the console, it opens over the scene rather than navigating away) that owns your preferences and your integration surface. It is organized into tabs: Display, Sim, Chat, Account, plus an API tab for signed-in developers and a Federation tab on enterprise plans.
The Settings panel on the API tab: a scoped token list with origin badges, and the four-stage connection test with per-stage latency results.
Display
Controls the globe's appearance:
- Earth style: Realistic, Operational (default), Dark, or Tactical.
- Quality preset: Cinematic (detail imagery and clouds), Balanced (detail imagery, no clouds), or Performance (neither); Performance is the default on touch devices.
- Layer toggles: links, labels, and sky glow.
Preferences persist in the browser per device, which is deliberate: the NOC wall display and your laptop want different quality tiers.
Sim
Defaults for the timeline: startup mode (live, replay, or demo), the default replay window, auto-play on boot, and default playback speed. These set how the console wakes up; the transport can override all of them live.
Chat
Shows the assistant's connection status honestly: live (production chat service), local (development backend), or offline (no backend; local intents only). If chat answers feel limited, this tab is the first thing to check; offline mode is labeled here rather than silently degrading. See Assistant.
Account
Your identity and plan: current subscription, query usage for guests, and plan management. Plan capabilities are covered in Permissions and tenancy.
API
The API tab is the integration surface for connecting the console to the platform, and it is honest about a platform that is still growing token infrastructure:
- Tokens. Create scoped API tokens (for example topology read plus telemetry write) for pipelines and the fleet agent. Each token is labeled with its origin: server-issued tokens come from the platform's token service, while browser-minted local credentials, generated when the token backend is not yet available for your account, are explicitly labeled local and will not authenticate against the platform. The console never lets a local placeholder masquerade as a live credential.
- Connection test. A staged probe that isolates failures instead of reporting one opaque error. Four stages run in order, each with its own pass/fail detail and latency: API reachability, the telemetry subsystem, the topology subsystem, and authenticated topology access with your key. A CORS problem, a dead subsystem, and a bad key each fail a different stage, so the fix is named by the test.
- Endpoints and headers. The tab shows the resolved API base URL and the header conventions your integration should use, matching API authentication.
Federation
Enterprise deployments with dedicated tenant stacks configure cross-stack federation here. The tab appears only on enterprise plans.
Common workflows
| Task | How |
|---|---|
| Tame GPU load on a laptop | Display, quality preset Balanced or Performance |
| Make the console boot into yesterday's replay | Sim, default mode and window |
| Diagnose a limited assistant | Chat tab, check for offline status |
| Wire up a telemetry pipeline | API tab, create a scoped token, run the connection test |
| Find out why live data is not flowing | Connection test; read which stage fails |
Example scenarios
A failed integration, diagnosed in one screen. A developer's pipeline posts telemetry but the globe stays empty. The connection test passes reachability and the telemetry subsystem but fails authenticated topology access: the token in the pipeline was a browser-minted local credential, clearly badged as such in the token list. Creating a server-issued token and re-running the test turns all four stages green.
Setting up the wall display. For the NOC monitor, an operator sets Display to Operational at Cinematic quality, Sim to live mode with auto-play, and leaves chat signed out. The preferences persist on that machine independently of everyone's laptops.