Airfy Compass
Manage every subscriber's WiFi from one screen.
Compass is the operator side of Airfy. Multi-vendor visibility, AI auto-resolve, and a real LLM interface for your support team. Built for internet service providers managing thousands to millions of homes.
The same backend running in production for hotels, banks, schools, and manufacturing today.
What Compass shows
The operator console you actually use every day.
Live, not by email digest.
Compass streams every event the network produces, filtered for the things that hurt revenue.
The reason ISPs ship platforms like this.
The category was created by competitors with documented numbers across millions of homes. Compass meets those numbers without the hardware rollout.
Industry baseline (Plume TCO Study, 2M+ subscribers)
Same study; auto-resolve closes most home-WiFi tickets without dispatch
No hardware rollout. Compass connects to the routers you already shipped.
Eight capabilities. One operator console.
Compass collapses the operator-side of WiFi management into a single platform: visibility, auto-resolve, an LLM interface, multi-vendor support, network safety, multi-tenancy, regional residency, and the scale to grow into.
End-to-end visibility
Per-subscriber diagnostics in real-time. Signal quality, channel utilization, device list, dropouts, mesh-link health. One screen, every home.
AI auto-resolve
Compass agents detect dead spots, slow speeds, mesh-link drops, and forgotten passwords. Most issues close themselves before a ticket is opened.
LLM operator interface
Your support team asks Claude or GPT in plain language: "Why is the WiFi slow at the Schmidt house?" Compass answers with a diagnosis, not a dashboard.
Multi-vendor
Manage Airfy nodes, consumer routers, telco-issued CPE, and incumbent enterprise gear through one operator console. Bring the hardware you already ship.
Network-level safety
Always-on threat blocking, malicious-site filtering, compromised-device quarantine. Enforced at the radio, not at the endpoint.
Multi-tenant by default
Customer hierarchies, partner sub-tenants, B2B2C white-label. Your CRM owns the relationship; Compass owns the network.
Where the data lives
Frankfurt and Austin regions today. EU residency for GDPR. SOC 2 in progress. Open source firmware so your audit team can verify what runs on subscribers' routers.
Built for scale
Same backend that runs in production for hotels, banks, schools, and manufacturing today. Hundreds of locations now, ten million ready when you are.
Compass under the hood.
Three layers, one platform. The cloud holds the ops dashboard, the analytics, the multi-tenant CRM hooks, and the LLM/MCP gateway. The edge runs on subscribers' existing routers (any modern subscriber router) and on Airfy nodes when you ship them. Between the two, agentic networking watches every link in real-time and acts before subscribers notice.
- Cloud: airfy.ui operator console, ISP CRM hooks, multi-tenancy, audit log
- Agent layer: MCP tools, network graph, anomaly detection, auto-resolve playbooks
- Edge: standard remote-management + vendor APIs, optional NPU for local inference on Airfy nodes
- Identity layer: per-subscriber, per-device, per-app rules — the same primitives that power the consumer app
┌──────────────────────────────────────────────┐
│ Compass — operator console (airfy.ui) │
│ dashboards · alerts · auto-resolve · CRM │
└─────────────────────┬────────────────────────┘
│ REST · GraphQL · MCP
┌─────────────────────┴────────────────────────┐
│ Agent layer (agentic networking) │
│ network graph · MCP tools · LLM gw │
└─────────────────────┬────────────────────────┘
│ Vendor APIs · Open standards
┌─────────────────────┴────────────────────────┐
│ Edge (subscribers' existing CPE) │
│ Consumer router · Telco CPE · Airfy node │
└──────────────────────────────────────────────┘
Why Compass is built differently
Four decisions that separate Compass from the incumbents.
Some platforms in this category sell ISPs a feature to create precision-marketing segments from subscriber behavior. Compass does not. Subscriber data exists to keep their network working.
Compass speaks every modern router protocol. Subscribers keep the CPE you already shipped. Rollout in 24 hours, not 18 months.
Frankfurt and Austin today. Edge processing on the S-class router for latency-sensitive workloads. On-prem possible for sovereign deployments.
MCP server included. Claude, GPT, Gemini, or your own agent can query the network in natural language. The dashboard is still there. The dashboard is no longer the only interface.
Bring your own hardware. Compass is the brain.
Compass talks to the routers your subscribers already have. No rip-and-replace. No 18-month rollout. Day one, every subscriber home shows up in the operator console.
For your engineering team
The operator console is one of many surfaces.
Your team probably wants the dashboard for support, the API for billing integrations, webhooks for alerting, and the MCP server for the AI agents you are already building. Compass ships all of them.
REST API
Subscriber, network, and device endpoints. OAuth2, rate-limited, versioned.
Webhooks
Outbound events for tickets, anomalies, firmware updates, SLA breaches.
MCP server
Model Context Protocol. Your AI agents query the network the same way Compass does.
Streaming logs
Live device-level events for ops teams. WebSocket + log broker.
GraphQL (beta)
For dashboards and reporting layers. Shipping Q3.
Postman + OpenAPI
Collections and specs published. Your SDK in a day.
Questions ISP teams actually ask.
How is Compass different from Plume Haystack?
Three big differences. (1) Compass is multi-vendor: it manages the routers your subscribers already have, not just our hardware. (2) Compass exposes a real LLM interface — your support team queries the network in natural language, not just dashboards. (3) Compass does not ship a marketing-segments feature. Your subscriber telemetry exists to keep their network working, not to build advertiser profiles.
Does Compass replace our existing routers?
No. Compass connects to the CPE you already shipped through remote-management protocols and vendor APIs, or any combination. Subscribers keep their hardware. Operators get a single console.
Where does subscriber data live?
Frankfurt for EU customers, Austin for North America. EU residency by default; we never move EU subscriber data outside the EU. SOC 2 audit in progress, ISO 27001 next. On-premise deployment is available for sovereign or regulated environments.
How long does deployment actually take?
First subscriber home onboarded the same day. A 100,000-subscriber pilot typically lights up in 30 days. Full ISP rollout in three months. Compare to a hardware-replacement rollout, which is measured in years.
What happens to our existing support team?
They stay. Compass deflects most tier-1 tickets, which means your team handles fewer angry calls and more interesting work. Net: lower headcount growth pressure, higher CSAT, faster resolution on the calls that do come in.
Can we white-label the subscriber app?
Yes. Your brand, your colors, your support flows. Compass runs underneath. Subscribers see your ISP, never see Airfy.
What about pricing?
Per-subscriber tiers, with volume discounts that kick in fast. Talk to the ISP team for a quote against your specific subscriber count, hardware mix, and rollout timeline.
Stop fighting tickets. Start running the network.
Compass is in production with operators today. We will give you a live demo against your own subscriber dataset, show you the auto-resolve queue draining itself, and put real numbers on what a 30-day pilot looks like.