Console overview
The Constellation OS console is the operator surface for your fleet. It runs in the browser at console.constellation.space/app/, renders your constellation on a photoreal globe, and puts every operational question, from "is this link closing" to "where should the next gateway go", one selection or one sentence away.
There is no page navigation. The console is a single scene: the globe fills the viewport, and every other surface (entity cards, the timeline transport, the assistant, panels) is an overlay on top of it. You never leave the scene to inspect an asset, run an analysis, or build a report.
Full console view: globe with the operator fleet, left rail, an open entity card on the right, sim transport at the bottom, and the assistant composer.
The four surfaces
Globe
Live situational awareness: satellites, ground stations, data centers, feeder and inter-satellite links, catalog constellations, and your own uploaded or synthetic assets, all rendered as selectable entities. Satellite motion is interpolated at display refresh rate, so the scene stays smooth at any playback speed. See Globe and rendering.
Timeline
One simulation clock drives everything. Play, pause, scrub, and change speed across a live feed, a stored replay window, or the bundled demo. Every metric on the globe, in entity cards, and in assistant answers refers to the timeline instant, not wall time. See Timeline and simulation.
Entity card
Select any node or link to open the card on the right: overview and posture, predictions, links, services, contact schedule, capacity, and incidents. Satellite selections show full orbital detail; gateway selections show elevation masks and the pass timeline. See Satellites and constellations and Gateways.
Assistant
A natural-language surface over the same scene. It reads the fleet you see, drives the UI (select, seek, camera, layers), runs deterministic analytics and mission-design engines, calls production prediction models, and returns auditable insight cards. See AI assistant and the agentic docs for the full capability catalog.
Alongside these, the left rail opens panels for catalog search, the chart gallery, the asset browser, reports, settings, and tips.
One scene, many sources
Everything on the globe lives in a single asset registry. Where an asset came from is metadata, never behavior: renderers, filters, analytics, and the assistant all consume an asset's capabilities, not its origin.
| Source | What it is | Telemetry | Links | Predictions | Orbital elements | Editable |
|---|---|---|---|---|---|---|
oneweb | Operator fleet (live and replay) | Live | Yes | Yes | Yes | No |
gsaas | Third-party ground station catalog sites | Live | Yes | No | No (ground sites) | No |
catalog | Catalog constellations (Starlink, GPS, ...) | None | No | No | Yes | No |
import | Files you uploaded | Recorded | Yes | No | If TLE-backed | Yes |
chat | Assets ingested through the assistant | Recorded | Yes | No | If TLE-backed | Yes |
synthetic | Generated what-if assets (mission sandbox) | Recorded | Yes | No | Yes | Yes |
Capabilities decide which analyses each surface offers. A TLE-backed upload gets the same SGP4 propagation, pass prediction, and coverage math as the operator fleet; a catalog constellation gets orbits and geometry but no telemetry. The provenance badge is the only difference you see.
How the pieces relate
- Selecting an asset on the globe opens its entity card and automatically pins it as context for the assistant, so "predict SNR on this link" needs no ids.
- The timeline clock is the single source of truth: the globe interpolates positions from it, entity cards sample telemetry at it, and every deterministic analytic reports the timestamp it computed against.
- Assistant analyses that produce charts or insight cards are captured into the Reports workspace with full provenance.
- What-if scenarios (outages, synthetic constellations, weather) mutate a copy of the scene deterministically and can be saved and reproduced. See Simulations and scenarios.
Access levels
| Capability | Guest | Signed in |
|---|---|---|
| Globe, timeline, demo fleet | Yes | Yes |
| Assistant | 10 trial messages | Unlimited (plan model) |
| Live fleet and replay telemetry | No | Yes |
| Predictions | No | Per plan tier |
| API tokens and integrations | No | Yes |
Plans and prediction tiers are covered in Permissions and tenancy.
A note on tenancy
Today the console is scoped to a single operator fleet per tenant: the operator program plus everything you upload, import through chat, or synthesize. Your imported assets live in your browser profile and optionally sync to your tenant's telemetry store; they are never visible to other tenants. Multi-organization workspaces and role-based access are on the enterprise roadmap, described honestly in Permissions and tenancy.
Example scenarios
A shift handover. An operator opens the console, glances at the posture bar, selects the one degraded gateway, reads its pass schedule and rain fade risk on the entity card, then asks the assistant for a daily fleet health report. The report lands in the Reports workspace with every number traceable to a tool run.
A capacity question from planning. A network engineer asks "where should our next three gateways go to maximize contact time". The assistant runs the ground network optimizer against propagated orbits, plays the candidate sweep on the globe, and returns a ranked table with each pick's marginal gain, all captured as a report they can export to PDF.
Where to go next
- New fleet? Start with Fleet setup.
- Bringing your own data? See Uploads and TLE ingestion.
- Connecting programmatically? See the API overview and Quick start.