What you can run on it

Six cases, each with what works today and what does not yet. The labels mean:

Available
works today
Pilot
being built with a first user or partner
Coming
next in line, not released
Planned
an idea, not available yet

Smart home & cameras

Available

Home Assistant · Frigate · MQTT

Camera snapshots and doorbell clips go to a cloud API to be tagged, with a bill per camera, and stop working when the internet does.

How it works with the gateway

  1. Home Assistant, Frigate or any script sends the snapshot to the hub with its own scoped key.
  2. The hub routes it to the node whose chip runs the model best, for example a TV NPU running EfficientNet-Lite4 int8 in 22 ms.
  3. Labels or an embedding come back; the image never leaves your network and every call is in the audit log.

Today and next

AvailableNative and OpenAI-compatible HTTP API, callable from Home Assistant automations and scripts today
ComingMQTT bridge (an example integration): a motion, doorbell or new-photo event triggers a model and the result is published back

Collector & scanner apps

Pilot

e.g. a trading-card recogniser

A card scanner needs a detector and an embedding model on a GPU server: the app maker pays for every scan, and it fails offline at a tournament.

How it works with the gateway

  1. The pipeline splits into two models: a rotated-box detector finds each card, and an embedding model plus a nearest-neighbour search over about 14k card arts names it.
  2. Both install as one package on the user’s own phone or TV; the small card index ships with it and updates with each card set.
  3. The app calls the user’s hub instead of its own server, so photos stay on the user’s devices.

Today and next

AvailableMulti-model nodes, image embeddings and classification through the hub
PilotPackaging the card detector and embedder as a gateway package (export to TFLite, accuracy checked against the server version)
ComingSigned, versioned packages distributed through the marketplace

Privacy-first consumer apps

Available

Android client library · same-device IPC

A photo app either uploads photos to embed them, or ships its own 582 MB model and loads it on every phone.

How it works with the gateway

  1. The app asks the client library for an embedding. With the gateway on the same phone, pixels go over Binder IPC in shared memory: no network, no JPEG loss.
  2. Away from the gateway, the library uses the home hub, then the app’s own bundled model as a fallback.
  3. Every path uses the same model and preprocessing, so the vectors stay comparable.

Today and next

AvailableSame-device IPC with a signature permission and a genuine-gateway check, plus hub and app fallbacks
AvailableIdentical results through the hub (cosine 1.000000); 0.996 against the app’s own on-device embedder
AvailableCLIP ViT-L/14-336 at ~1.45 s per image on a Pixel 7 Pro GPU

Client library docs

Office & edge fleets

Available

Keys · quotas · audit · private registry

Offices are full of screens, boxes and spare phones with NPUs, but nobody can manage them as one inference pool, or prove who used what.

How it works with the gateway

  1. One hub per site; every device joins as a node with a node key and dials out, so no firewall changes.
  2. Each team or app gets a virtual key with model scopes, a rate limit and a daily quota; every request is audited.
  3. The registry records which model runs on which of your devices, and the router uses it.

Today and next

AvailableVirtual keys, scopes, rate limits, daily quotas, audit log, usage, Prometheus metrics, admin dashboard
PlannedSSO, role-based access, audit export, a private registry and fleet management

Free, no paid tiers

Retail “AI showroom”

Pilot

Display TVs · in-store AI

Stores run dozens of always-on display TVs that only loop a demo video, while in-store AI (shelf checks, product recognition) needs separate hardware.

How it works with the gateway

  1. The display TVs become nodes and serve models live: “these TVs are running AI right now” is the demo.
  2. In-store apps call the store’s own hub, on hardware the store already owns.
  3. Display TVs run the maker’s retail demo mode, so this needs cooperation from the TV maker or the store chain.

Today and next

AvailableA consumer TV serving models on its NPU (measured on a TCL Google TV)
PilotA partnership with a TV maker or store chain: a natural OEM pilot

For device makers

Share your devices, maybe get paid

Planned

Planned, not available yet

Model makers cannot test a model on hundreds of real phones, TVs and boxes with different chips, firmware and NPUs. Owners of those devices get nothing for their idle hardware.

How it works with the gateway

  1. Opt in, and your idle devices run real-device benchmarks for model makers: “test my model on 500 real devices”, like a device farm, but for AI.
  2. Only while the device is idle (never while you watch), within the power and data budget you set; jobs are verified with spot checks.
  3. Planned, not available yet: whoever uses your devices could pay you, with a small platform fee. Not built, not priced, no dates; today sharing is free and only within trusted groups.

Today and next

AvailableThe pipeline it reuses: nodes, benchmarks and the registry
AvailableTrusted groups (free sharing with people you know), owner protections (idle-only, charging, thermal), end-to-end sealing
PlannedPayment between users, metering, result verification

What a shared device is worth

Honestly: not much per job. A TV is on a few hours a day and suspends shortly after it is switched off, and a datacenter GPU does bulk inference very cheaply. Sharing makes sense where a datacenter cannot help: work that must stay on real, local devices, and people who would rather use a friend's idle NPU than a cloud.

What a datacenter cannot offer is real-device diversity: thousands of different TVs, phones and boxes with different chips, firmware and NPUs. Testing a model on them could be useful to model makers, app developers and chip and TV makers. Planned, not available yet.