EdgeLens AI | On-Device Ambient Intelligence for the Connected Home
On-device ambient intelligence

Privacy-first ambient intelligence for the smart home.

EdgeLens AI helps smart-home and networking OEMs transform routers, cameras, and in-home sensors into on-device AI agents that understand presence, learn daily rhythms, and surface meaningful changes for families and caregivers — with raw video and sensor streams kept local by default.

ON-DEVICE AIMom — active 7:15Quieter than usualraw data stays localderived alerts out ↑
Demonstrated live at CES 2026 and NVIDIA GTC
Wi-Fi CSI sensing proof-of-concept completed
CPU-first router-class integration in progress
Guardian platform core built & under active validation
Founded by a former Amazon Alexa AI/Search product leader
Private by architecture — no raw home data leaves the device
Two sides of the home

One on-device platform. Two buyers it serves.

We sell to the OEMs who ship the hardware — and we are designed around the people living with it.

Primary buyer

Smart-home & networking OEMs

Turn installed routers, gateways, and cameras into a premium, privacy-first home-awareness subscription — no cloud video, no new hardware, no NPU dependency.

  • On-device SDK / API + reference app; drop-in for firmware teams
  • Wi-Fi CSI + camera/radar fusion tuned for CPU-first silicon
  • Recurring software revenue, not per-inference cloud cost
  • Paid design-partner / NRE path to per-device licensing
Whom it protects

Aging-in-place & family care

Families and caregivers get quiet peace of mind — knowing a parent is okay without cameras in private spaces, wearables to forget, or a stream of clips to watch.

  • Presence, daily-rhythm baselines, and meaningful-change alerts
  • Wi-Fi sensing covers private spaces cameras can’t
  • Graded, calm signals — not alarm fatigue
  • A channel into care operators, ISPs, and AgeTech partners
Architecture · engineered for the edge

Built to run where the data is born.

The hard part isn’t a cloud demo. It’s private, reliable, low-cost inference on devices with tight CPU and memory budgets. EdgeLens is CPU-first by design, with on-demand multimodal verification — so the device’s primary job is never at risk.

DEVICESON-DEVICE AIFAMILY / CAREGIVERRouter / gatewayCameraIn-home sensorsWi-Fi CSI sentinelEdgeLens NanoML + fusionBehavioral baselinePrivacy-policy engineDerived contextGraded alertsno raw video
On-device footprint (design targets)
Baseline sensingCPU-first
NPU / GPUoptional
Always-on core≤ 1 pinned core
Raw video off-devicenever
Works offlineyes

Designed to port across router, camera, and hub silicon families via a thin integration layer.

SDK / APIreference appOEM-agnostic adapters
Flagship product

EdgeLens Guardian

A private, on-device home-awareness layer for aging-in-place. Guardian learns a home’s normal rhythm in days, then surfaces calm, graded signals only when something genuinely changes — routine drift, prolonged inactivity, or a bathroom overstay.

TIER 01

Awareness

Presence, home/away, daily-activity summary, and privacy zones.

TIER 02

Care

Behavioral baseline, meaningful-change & inactivity alerts, caregiver notifications.

TIER 03

Safety Premium

Fall-like event escalation with on-demand multimodal verification.

Graded caregiver signals — calm by default
GreenNormal rhythm
YellowRoutine drift
OrangeCheck-in suggested
RedAct now

EdgeLens Guardian is a home-awareness and wellness product. It is not a medical device or emergency-response system.

Technology

Wi-Fi sensing first. Cameras only when it matters.

Stage 1 · always-on

Wi-Fi CSI sentinel

Presence, occupancy, and 7-day behavioral baselines from the always-on router — covering rooms cameras can’t, pinned to a single CPU core.

Models

CPU-first small models

Purpose-built, quantized EdgeLens NanoML models tuned to run in real time on constrained home silicon.

Stage 2 · on-demand

Multimodal verification

Camera/radar fusion triggers only for higher-confidence events, then idles — protecting privacy and the device’s primary job.

The cloud-dependent way

  • Streams raw video off-device for analysis
  • Per-use inference cost grows with adoption
  • Latency, breach exposure, regulatory risk
  • Can’t cover bathrooms or bedrooms

The EdgeLens way

  • Processes locally; raw streams stay in the home
  • No per-inference cloud bill — software margins
  • Real-time, works offline, private by design
  • Wi-Fi sensing reaches private spaces, camera-free
Why now

The OEM design window is open.

On-device small models can finally run in real time on constrained hardware. IEEE 802.11bf-2025 standardized Wi-Fi sensing. Aging demographics and a caregiver crisis make private home awareness urgent — while rising cloud-AI costs squeeze OEM margins. The teams that embed on-device intelligence in this window will define the category.

FAQ

Questions buyers ask first.

What is EdgeLens AI?+

EdgeLens AI builds privacy-first, on-device ambient intelligence for the connected home. Our software turns routers, cameras, and in-home sensors into local AI that understands a home’s daily patterns and surfaces meaningful changes, without sending raw home data to the cloud. We sell to smart-home and networking OEMs.

Who is it for?+

The buyer is smart-home and networking OEMs — router, gateway, camera, and hub makers. The end beneficiaries are families and caregivers monitoring aging parents.

How is it different from cloud camera AI?+

Cloud camera AI streams video off-device — adding cost, latency, and privacy risk — and it can’t cover private spaces. EdgeLens runs on-device, uses Wi-Fi sensing to cover those spaces, and keeps raw video and sensor streams local by default.

Does EdgeLens send video to the cloud?+

No. Raw video and sensor streams stay local by default. EdgeLens sends derived context, summaries, and authorized caregiver alerts — not raw private home data.

What hardware does it run on?+

It is designed for CPU-first baseline sensing on constrained home silicon (router, camera, hub families), with optional camera/radar verification where device resources allow.

Why we’re building this

“The next wave of home AI can’t stay tethered to the cloud. We’re putting that intelligence in the room — quietly watching for what matters, for the people you love.”

NL
Nagendra LavuFounder & CEO · ex-Amazon Alexa AI/Search · ex-Cerence VP/GM
Stage
CompanyPre-seed
WedgeAging-in-place
ChannelNetworking OEMs
Based inSan Jose, CA

Build private home intelligence into your next platform.

Engaging select smart-home, networking, and AgeTech OEMs for paid design-partner evaluations.