Read-only demo. A simulated multi-vendor fleet. Detection and fixes are illustrative; nothing here touches a real network.

Settings

Organizations and team, vendor connections, the kill switch, the confidence gate, and notification routing.

Safety

Kill switch: released · RCA threshold: 70%

Adjusting these is disabled in the read-only demo.

Vendor connections

LAN probe (this host)
On-LAN collector · ping / DNS / Wi-Fi
Active

Runs in-process on this machine, sampling the home network continuously (uplink reachability, DNS health, Wi-Fi signal) and feeding real faults into the same pipeline. Consumer routers have no write API, so most probe faults are detect-and-triage; DNS-cache issues get a safe one-tap flush.

Meraki Dashboard
Cloud-to-cloud · webhooks + Dashboard API
Not connected

API key stored in Supabase Vault and decrypted only by the server-side executor at fix time. Without an active credential, executing a Meraki fix is blocked.

UniFi Controller
On-LAN · local collector
Not connected

Store controller credentials now; the live adapter ships next pass. The canonical model is already vendor-neutral, so wiring it in is additive.

Notification routing
Slack
#net-ops (proposal + approval actions)
Off
Microsoft Teams
NetOps channel
Off
MSP webhook
POST https://… (external console)
Off

Channel delivery lands with the secrets/notifications pass. The in-app pane is always the source of truth, so an unreachable channel never loses a fault.

Data export

Export the fleet state or the full audit trail without developer involvement.