The Artificial Intelligence of Things

Big thinking. Home-sized.

AI can have a conversation. Your home can have a hundred connected things. ALYT brings them to the same table.

See it come togetherFind your hub

Sense.Connect.Understand.

Follow the signals

Four letters. A very different home.

So, what is AIoT?

AIoT stands for the Artificial Intelligence of Things: artificial intelligence combined with connected devices. Sensors and devices provide information; AI helps interpret requests and context; connected controls turn supported requests into action.

ALYT brings that idea home with Simone, our custom-trained home assistant, a shared understanding of rooms and device capabilities, and routines that run on the hub. You choose what is connected and how your home responds.

Sensor fusion · the next possibilities

Small signals. Bigger picture.

A motion detector notices movement. A contact sensor notices an open window. A power meter notices a change. Put those clues together, and ordinary sensors could tell a much richer story. That is sensor fusion—and a compelling role for Simone’s local AI.

These are potential AIoT experiences, illustrated as concepts. They depend on supported sensors, integrations and analysis; they are not a list of currently available features.

Potential experience

Your thermostat has a second calling.

The sensor that helps manage heating could also contribute to a clearer picture of activity. Combine its motion or occupancy reading with a recent door opening and radio changes from a sensing-capable router integration. Suddenly, there is more to go on.

The clues

Thermostat sensorMotion or occupancy
Router sensingRadio changes
Door contactRecent opening
SimoneLocal AI on Hub AI

A possible interpretation

Activity worth a closer look

Simone could explain

“There are several signs of activity in the living room while Away is set. Here is what contributed.”

How the clues add up

Simone could compare the timing and location of several clues, and how fresh each one is. Together, they could strengthen a presence estimate or highlight activity while the home is set to Away. An estimate stays an estimate; it does not identify a person or prove an intrusion.

Radio-disturbance sensing needs suitable hardware and integration support. A phone joining Wi-Fi or a signal-strength change alone cannot establish room occupancy.

Potential experience

Heating the room. Or the neighbourhood?

The window is open. The room is cooling. The heating is still working. A contact sensor, a temperature reading and the thermostat’s operating state could turn an unexplained cold room into an explanation you can use.

The clues

Window contactOpen
Room temperatureFalling
Heating stateStill active
SimoneLocal AI on Hub AI

A possible interpretation

Heat may be escaping

Simone could explain

“The room began cooling after the window opened, while the heating stayed on. Shall we adjust it?”

How the clues add up

Simone could compare the temperature trend before and after the window opened, alongside heating activity. That context could support a timely suggestion to close the window or adjust heating, with any change following the household’s permissions.

Requires compatible window, temperature and heating-state inputs. An explanation is a hypothesis, rather than a diagnosis of the heating system.

Potential experience

An ordinary appliance. An extra sense.

A washing machine does not need its own AI to contribute useful clues. A compatible power-monitoring plug and a vibration sensor could let Simone interpret its changing power use and movement over the course of a cycle.

The clues

Power monitorBack near idle
Vibration sensorMovement settled
Recent historyCycle pattern
SimoneLocal AI on Hub AI

A possible interpretation

The cycle may be complete

Simone could explain

“The power and movement patterns suggest the washing cycle has finished.”

How the clues add up

Local analysis could compare the sequence over time, including normal pauses, to estimate when the cycle has finished. It could offer a useful heads-up through a supported local interface, without replacing the appliance or sending that analysis to a remote model.

Requires suitable sensors and validated analysis for the appliance. This example concerns a notification, with no automatic interruption of appliance power.

Edge AI · intelligence at home

Internet out. Intelligence in.

Hub AI runs Simone’s model in the home. With compatible local inputs, reasoning can stay close to the room—even during an internet outage. Local routines and sensor-based protection run on every ALYT hub; Hub AI adds onboard model reasoning.

Internet unavailable

Inside your home · powered and locally connected

  1. Local sensors
  2. Hub AI
  3. Local response
Architecture illustration. Cloud-only inputs and services remain unavailable during an internet outage.

Skip the cloud commute.

Local analysis removes the internet round trip to a remote model. The decision stays close to the devices that supply the clues.

Keep the context close.

Onboard reasoning lets supported local sensor data be analysed in the home. Cloud integrations and any separately authorised data sharing keep their own requirements.

Keep thinking through the outage.

A working local sensing and control path can continue when the internet connection fails. The hub, sensors and the local connections they use still need power and connectivity.

Several ways to stay connected.

Compatible Zigbee and Bluetooth LE devices can talk directly to the hub without your Wi-Fi router. Matter over Thread uses a local mesh; Wi-Fi devices use the home network. Local operation avoids dependence on an internet connection, while each device still needs its own working local path.

Thread uses IPv6, and Wi-Fi/LAN integrations use IP networking. Local AI means no cloud round trip for local model processing; it does not mean every device avoids TCP/IP.

Meet Hub AI

A little conversation. A real-world response.

Good thinking has a ripple effect.

Follow a request from you to your home. Or let a routine you set do the remembering.

Your connected home

Living roomLights · 30%
Your requestOne room
Your choiceOne setting
“Dim the living room lights to 30 percent.”

The home responds

A few words. A softer evening.

Simone turns your request into a control for the compatible lights in the room you named. The percentage is yours. The extra taps can take the evening off.

  1. You say it

    Name the room and the change you want.

  2. Simone interprets

    The request is matched to supported device controls.

  3. ALYT acts

    The command goes to the compatible lights.

Natural-language control · compatible dimmable lights

Your connected home

Selected lightsSwitch off
CurtainsClose
Saved routineYour sequence
“Run Good Night.”

The home responds

One good night. Several good decisions.

Call a routine you have already saved. It can switch off selected lights and close compatible curtains in the order you chose. You wrote the plan; the hub runs it.

  1. You ask

    Call your saved Good Night routine.

  2. Simone selects

    Your named routine supplies the actions.

  3. The hub runs it

    Local steps run at home; online steps use their services.

Saved routine example · configure the devices and actions first

Your connected home

SunsetScheduled
Local routineOn the hub
Local lightsSwitch on
Sun goes down. Lights come up.

The home responds

The sky sets the time. You set the mood.

Set a sunset routine for compatible local lights. The hub runs the saved rule at home. This is local automation doing its job—even when a cloud conversation is unavailable.

  1. Sunset arrives

    The home’s location sets the schedule.

  2. Your rule matches

    The hub checks the conditions you saved.

  3. Local lights respond

    Local execution needs no cloud AI decision.

Local automation · the hub and devices must remain powered and connected

Interactive illustrations of supported capabilities. Actual responses depend on your connected devices and saved routines.

The clever part is how it fits together

Brains. Connections. Follow-through.

An assistant with a home life.

Simone is ALYT’s custom-trained home assistant. Ask naturally about your rooms, supported devices and saved routines. The model handles language; the platform supplies the controls.

Different brands. Shared understanding.

Connectors describe devices through common capabilities. A light is still a light, even when it speaks a different protocol. Wi-Fi, Bluetooth LE, Zigbee and Matter over Wi-Fi or Thread connect compatible devices.

The everyday stays close.

Local routine execution and sensor-based protection live on every hub. Cloud-connected devices and services still need their online connection. Local AI reasoning is the additional choice offered by Hub AI.

Know where the thinking happens

Cloud or couch-side. Your call.

Every model brings your devices together and runs local routines. Choose where Simone’s model does its reasoning, and the sound you want beside it.

ALYT Hub

Cloud Simone. Local routines.

Use compatible connected speakers for sound.

ALYT Hub Music

Cloud Simone. Built-in sound.

Local routines, with a full-range speaker on board.

ALYT Hub AI

Onboard Simone. Local reasoning.

Local routines, with premium built-in sound.

Local reasoning does not make every service offline. Music streaming, calendars and cloud device integrations may still need internet. Cloud AI requires household authorization.

Compare all three hubs

A little substance behind the sparkle

Curiosity welcome. Details included.

Explore the product architecture, compatible connections and documented workflows. The good questions deserve more than a tagline.

Hub specifications

Model-by-model features and connection options.

Compatible connections

See what connects, and how.

Routines in practice

Read the homeowner guide to saved routines.

Build with ALYT

Explore the public API, SDKs and connector documentation.

Good questions. Straight answers.

How is AIoT different from ordinary home automation?

Home automation executes rules, such as turning on a light at sunset. AIoT brings AI into connected-device systems: interpreting language and, with supported analysis, drawing meaning from several sensor signals. ALYT combines Simone with explicit controls and saved routines; the sensor-fusion stories above illustrate further potential. A scheduled rule does not need a fresh AI decision.

Does an ALYT home work without internet?

Local routines and compatible local device controls run on the hub while it and the devices remain powered and connected. Cloud AI on Hub and Hub Music needs internet. Hub AI runs Simone’s model locally, but online content and cloud integrations still need their services.

What is the difference between local AI and cloud AI?

Cloud AI sends a request to an authorized online AI service for model processing. Local AI runs the model on hardware in the home. ALYT Hub and Hub Music use cloud Simone; ALYT Hub AI hosts Simone’s model on the hub. Local routine execution is included on all three.

Is Matter the same thing as AIoT?

No. Matter is a standard for communication between compatible smart-home devices. AIoT describes combining AI with connected things. ALYT supports Matter over Wi-Fi or Thread on all three hub models, alongside Wi-Fi, Bluetooth LE and Zigbee connectivity.

Can ALYT bring different device brands together?

Yes, through supported connectors and shared device capabilities. Compatibility depends on the specific device and connection. Check ALYT’s connector catalogue for supported brands and workflows; a shared radio standard alone does not mean every feature of every device is supported.

Does Simone invent routines or change the house on its own?

The control examples use explicit requests and saved routines. The sensor-fusion concepts illustrate how supported analysis could turn observations into useful explanations or suggestions. Actions must follow the household’s permissions. These concepts do not promise autonomous security decisions or changes without permission.

What is sensor fusion in an AIoT home?

Sensor fusion combines observations from multiple sources, such as motion, temperature, door contacts and power use. Considering their timing, quality and context can support a more useful interpretation than one reading alone. The examples on this page are potential experiences; each needs supported inputs and validated analysis.

Can every router detect movement in a room?

Router integrations can report connected-device presence and, where available, signal strength. Detecting movement from changes in radio propagation needs suitable sensing hardware, measurement access and analysis. Those are different capabilities. A Wi-Fi connection or a change in signal strength alone is not reliable proof that a person is in a particular room.

Does edge AI mean the home stops using TCP/IP?

Edge AI describes where the model runs: on local hardware. It can operate without an internet connection when its inputs and outputs are local. Zigbee and direct Bluetooth LE connections can bypass a Wi-Fi router; Thread uses IPv6, and Wi-Fi devices still use IP networking. Local connections must remain operational.

A home with more on its mind. And you with less on yours.

Register your home, connect what you already own, and put the hub in the room where you talk the most.

Register your home Meet the hub