← Beau Sumile
Prop‑eye working name
Edge‑AI outdoor vision for property management

See what's happening at your property —
without sending the footage anywhere.

Prop‑eye is an edge‑AI outdoor camera system that delivers property‑management‑grade intelligence — party formation, vendor accountability, damage detection, portfolio analytics — without the cloud‑recording privacy problems of Ring or Nest. All inference runs on a small box at the property; only event summaries ever leave the network.

Outdoor‑only · sidesteps Airbnb's indoor ban Edge‑first · raw video never leaves the property 9 higher‑order event classifiers $35–65/property/mo $500M–$1B TAM at maturity
01 — Purpose

Why Prop‑eye exists

Existing property monitoring splits into two camps: privacy‑safe noise sensors that are blind to visual events, and consumer cameras that have no idea what a property manager actually cares about. Prop‑eye sits in the gap — visual intelligence that speaks property‑management semantics, built privacy‑first so it survives a lawyer review and a platform policy.

What it is

  • An outdoor‑only edge‑AI vision system
  • Local processing — raw footage never leaves your network
  • A property‑management event model (party formation, vendor verification, damage detection)
  • A platform that scales from 1 property to N
  • A genuine answer to “what's happening at my property right now?”

What it is not

  • An indoor camera system (Airbnb‑banned, not the goal)
  • A face‑recognition database — re‑ID by appearance, not identity
  • A replacement for monitored alarm services
  • A consumer Ring/Nest replacement
  • A pure security product — it's operational intelligence with security as one event class
02 — The flow

How an event travels through Prop‑eye

A single detection fans out across nine parallel classifiers, each scoring what it sees, then converges through severity triage to a single notification — all of it on‑device. The only thing that crosses the edge boundary to your phone is a summary and a thumbnail.

Capture
Front‑door camRTSP / ONVIF
Side‑yard camRTSP / ONVIF
Street‑view camRTSP / ONVIF
Detect
Frigate NVRYOLOv8/v11 · Jetson GPU
Stream
MQTT event busMosquitto
Normalize
Event processornormalize · dedupe
Classify
Vendor arrival
Unknown visitor
Party formation
After hours
Loitering
Kid safety
Package delivery
Damage detect
Yard activity
Route
Severity routerscore 1–10 · idempotent
Triage
Silent logsev 1–3
Standard pushsev 4–6
Urgent pushsev 7–9
Criticalsev 10
↓ summaries + thumbnails only — no raw video crosses the edge ↓
Owner's phone & dashboardntfy · Pushover · <30s
brand / edge‑local path logged only (sev 1–3) standard push (4–6) urgent push (7–9) critical escalation (10)
03 — Market fit

Where Prop‑eye lands

Three customer segments with different pain, different willingness‑to‑pay, and different competitive dynamics. The opportunity isn't “AI cameras for STRs” — Airbnb's 2024 indoor ban killed that. It's outdoor visual intelligence for the operators who manage doors at scale.

Segment A

STR single‑property owners

~15M US units. Largest by count, smallest by dollar‑per‑customer. Parties, damage disputes, unauthorized guests. Heavily price‑sensitive and already saturated with noise monitors.

Willingness to pay: $10–30/mo/property
Segment B

STR property managers

10–500 doors at 15–25% management fees. The juiciest segment: crew accountability, maintenance detection, insurance documentation, a portfolio dashboard their PMS lacks. They pass costs through to owners.

Willingness to pay: $25–75/mo/property
Segment C

Long‑term rental investors

~24M US individual‑landlord units. Smaller by count, stickier and higher value per door. Condition verification, lease‑violation evidence, move‑in/out documentation — and no Airbnb‑style platform ban to worry about.

Willingness to pay: $20–50/mo/property

The wedge: No one offers outdoor‑only AI vision purpose‑built for property managers with portfolio‑level intelligence. Consumer cameras are single‑property; PM monitoring tools are sensor‑based and blind to visual events. That empty space is the opening.

“The portfolio intelligence layer for property managers — outdoor AI vision that pairs with your existing noise monitors to catch what they miss, while keeping all footage local.”

Competitive positioning

Cost / complexity vs. property‑management visual intelligence. The top‑center is unoccupied — that's the wedge.

OPEN WEDGE Cost / complexity → PM visual intelligence → Noise sensors Consumer cams PMS tools Enterprise (Verkada) Prop‑eye
04 — Features vs. competitors

What Prop‑eye does that the others don't

Noise sensors own the customer's mental model but are blind to anything visual. Consumer cameras have no property‑management awareness. Enterprise systems are priced for offices. PMS platforms manage bookings, not what's physically happening on‑site. Prop‑eye is the only one built for visual PM intelligence, privacy‑first.

Capability Prop‑eyeoutdoor AI vision Noise sensorsMinut, NoiseAware Consumer camsRing, Nest, Eufy EnterpriseVerkada, Rhombus PMS toolsGuesty, Hostaway
Outdoor visual intelligence
Party formation (before the noise)
Cleaning crew / vendor accountability
Damage / maintenance detection
Privacy‑first (no cloud video)
No face‑recognition database
Airbnb‑compliant (outdoor‑only)
Multi‑property portfolio dashboard
PMS integration (Guesty / Hostaway)
Works through an internet outage
Purpose‑built for property management
Price / property / month$35–65$10–15$3–10 +dev$100–300/cam% of rent
full partial none / not applicable
05 — Pricing

Three tiers, benchmarked to the market

Priced between sensor‑only monitors ($15) and enterprise cameras ($100–300/cam) — the midpoint that AI vision for property management actually warrants. BYOC (bring your own RTSP camera) for the first cohort; a recommended hardware bundle once the support load is understood.

Solo operator
$35 /mo/property
  • Up to 2 outdoor cameras
  • Standard event types
  • Mobile app + email alerts
  • 30 days of event history
  • Self‑install, no contract
Property manager
$65 /mo/property · 10+ door discount
  • Up to 4 outdoor cameras
  • Multi‑property dashboard + cross‑portfolio analytics
  • PMS integrations (Guesty, Hostaway, OwnerRez, AppFolio)
  • Cleaning crew + contractor workflows
  • Custom event types · 90‑day history
  • Phone support
Enterprise · 50+ doors
from $50 /mo/property
  • API access for custom integrations
  • White‑label dashboard option
  • Insurance claim documentation packages
  • 365 days of event history
  • Dedicated account manager + SLA
06 — The privacy moat

Defensible by default

Privacy isn't a feature bolted on — it's the architecture. It's also the moat: it survives a lawyer review, sidesteps Airbnb's platform policy, and is the thing incumbents can't easily copy without rebuilding.

What we collect

  • Detection events: timestamps, zones, class labels
  • Short event clips, auto‑purged after 14 days
  • Vehicle make / model / color
  • Rotating appearance embeddings (not permanent)
  • Vendor schedules and patterns

What we never do

  • Persistent face‑recognition database
  • Audio recording of any kind
  • Continuous 24/7 cloud footage
  • Store license‑plate text as PII
  • Point cameras at neighbors or public space

Hawaii posture: outdoor entry points aren't “private places” under HRS §711‑1111, and one‑party audio consent is moot because nothing records audio.

07 — Roadmap

From your house to twenty doors

Current state: pre‑hardware dev mode — 9 classifiers, 46 tests passing, 38 simulator scenarios

Built at the home testbed first. The discipline: don't enter the commercial phase until 60 days of honest, real‑world evaluation says it's worth it.

Phase 0
Foundation
Jetson, Docker, networking, auto‑recovery
Phase 1
Camera baseline
RTSP streams + Frigate detection
Phase 2
Enrollment
Family, vehicles, recurring vendors
Phase 3
Higher‑order events
The 9 classifiers + severity routing
building
Phase 4
Mobile
Phone alerts + live state page
Phase 5
Hardening
Health, backups, outage recovery
Phase 6
60‑day eval
Live with it, journal, decide
Phase 7
Commercial pilot
First PM doors + portfolio dashboard
08 — Risks to watch

What actually kills this

Airbnb extends the ban outdoors

Would gut the STR TAM. Mitigation: dual‑market into long‑term rentals from day one.

Incumbents add cameras

Minut/NoiseAware have the relationships. ~12–18 month lead; win on PM‑specific depth.

Biometric privacy regulation

BIPA‑style laws expand. Mitigation: design for the strictest jurisdiction; privacy as a feature.

Liability for missed events

Positioned as operational intelligence, not security service. Strong contract limits + E&O insurance.

Hardware reliability

Cameras die, Jetsons overheat. BYOC defers risk; remote diagnostics; realistic SLAs.

Founder bandwidth

The most likely failure mode. Solo CPA, two kids, parallel projects. Side project = it doesn't ship.