Skip to main content

Use the assistant

The assistant is how operators query, visualize, and forecast over the fleet without leaving the console. It receives the same topology, telemetry, and timeline state as the globe on every message.

Capabilities

CategoryExamples
AnalyticsP90 latency, utilization rollups, histograms by node type
Globe controlSelect nodes and links, pause or seek the timeline
PredictionsSNR fade, weather impact, congestion, RF jamming, collision risk
AssetsImport fleet files, list catalog entries
IntegrationsReturn API bundles for your IDE, agent, or CI

Example prompts:

  • “What is the P90 link latency at this instant?”
  • “Predict SNR on the selected link for the next 15 minutes”
  • “Chart gateway throughput for the replay window”
  • “Give me the integration bundle for Cursor”

Models and ML tiers

PlanAssistantPredictions
StandardConstellation StandardBronze tier
ProConstellation ProBronze, silver, gold
EnterpriseConstellation EnterpriseAll tiers + custom models

Unlocked tiers appear under Settings → Account. Prediction tools call the same Predictions API that edge automation uses—results are model output, not placeholders.

Connection status

Settings → Chat reports model label and backend connectivity (Connected, Local dev, or Offline for chart-only mode).

IDE and automation

Signed-in users can request an integration bundle—scopes, endpoints, install script—for IDE & agent handoff.