Smart Kitchen Devices That Actually Work With Home Assistant

This post may contain affiliate links. As an Amazon Associate we earn from qualifying purchases. Disclosure.

The kitchen is where smart home plans usually stall. Every other room fills up with devices that speak Zigbee or Matter and land in Home Assistant within minutes, and then you buy an oven for two thousand dollars and discover it talks to exactly one app, forever. So which appliances genuinely open up, and what do you do about the ones that don't?

TL;DR: Bosch and Siemens (Home Connect), Miele, LG ThinQ, and Samsung via SmartThings have official Home Assistant integrations in 2026, and every one of them runs through the maker's cloud. Samsung's free SmartThings API access ends in October 2026. Everything else is either cloud-locked or app-only. For those, energy-monitoring smart plugs plus door and temperature sensors cover most of the useful automation for a fraction of the cost, and they work on appliances with no smart features at all.

Why Is the Kitchen Harder Than Every Other Room?

Lighting, heating, and security all got commoditised. A smart bulb is a 12-dollar object that speaks a public protocol, and if the brand disappears, you replace it. A dishwasher is an 800-dollar object you expect to keep for most of a decade, and the manufacturer knows you're not swapping it because the app got worse.

That asymmetry shapes everything. Appliance makers build their connectivity as a customer-retention feature, not as an integration point, so the app is the product and the API is an afterthought. Home Connect is the closest thing to an exception, because Bosch and Siemens publish a documented cloud API with an open developer portal.

There's a second problem nobody warns you about: kitchens are hostile to radios. Metal appliance bodies, a microwave oven that works in the same 2.4 GHz band as Wi-Fi every time someone reheats coffee, and dense cabinetry between the router and the far wall. If devices keep dropping in one specific room, the kitchen is a prime suspect. Zigbee and Thread share that band, so they aren't immune, but mesh routing lets traffic go around a dead spot, which is a good argument for putting your sensor layer on one of them and picking a proper coordinator rather than a USB stick jammed behind the server.

Which Appliance Brands Actually Integrate?

Here's the state of play in 2026, based on Home Assistant's integration documentation and the manufacturers' own developer portals. I haven't tested these appliances myself, so the table lists what each integration documents rather than my own measurements.

Home Connect deserves its reputation. The integration exposes the operation state, finish time and program progress as proper entities, which means "notify me when the dishwasher finishes" is a short automation rather than a project. Setting it up takes more than adding a login, though. You register an application on the Home Connect developer portal and paste its client credentials into Home Assistant. The catch is that everything routes through the Home Connect cloud, so an outage takes your automations down with it. That's the trade you accept for the depth.

The cloud also comes with documented limits. Home Connect's API reference caps an application at 1000 requests per day and 50 per minute for each user account, and at 5 program starts per minute. Home Assistant listens to server-sent events instead of polling, which is the approach the API documentation itself recommends, so a normal setup stays far below those numbers. A script that polls state in a loop will not.

Samsung's route through SmartThings works, but adds a second cloud in the chain. If you already run SmartThings devices, the SmartThings integration brings appliances along with everything else, which is convenient. It's also two vendors between you and your dishwasher, and the terms are changing. Samsung announced on 23 June 2026 that free SmartThings API access moves to a paid plan from October 2026, at 4.99 dollars a month for individual developers, and Home Assistant's own documentation carries a warning about it. The SmartThings app itself is not affected, only things that call the API, which includes this integration.

Miele used to be the frustrating one, and it has improved. Home Assistant now ships an official Miele integration built on Miele's third-party API, with the same top quality rating as Home Connect. The limits are the usual ones. The appliance has to be connected to the Miele cloud through the Miele app, the API does not fully match the app, and the integration's documentation notes that endpoints for the newest washer-dryer and vacuum generations lagged behind the rest.

What About Matter? Didn't That Fix This?

Appliance device types arrived in two steps. Matter 1.2, released in October 2023, added refrigerators, dishwashers and laundry washers. Matter 1.3, released in May 2024, added ovens, cooktops, extractor hoods, microwave ovens and laundry dryers. On paper that's the answer. In practice, a device type only defines what a maker can expose, not what it will. I couldn't find evidence of many shipping appliances that implement these types, and a Matter implementation can be a shallow slice, such as on/off plus a cycle state, rather than the full control surface the native app has.

The device types themselves are public, and the Connectivity Standards Alliance specification is the place to check what an appliance category is even capable of exposing before you trust a marketing claim about it. The CSA's release notes for Matter 1.2 and Matter 1.3 list which appliance categories landed in which version.

I'd still take Matter over a cloud integration where the choice exists, because Home Assistant's Matter integration works locally with no vendor cloud involved, and that survives the manufacturer losing interest. Just don't buy an appliance on the strength of a Matter logo and expect parity with Home Connect. If the standard itself is new to you, the Home Assistant setup guide covers where Matter fits alongside the older protocols.

What Is the Cheap Layer That Works on Everything?

This is the part that matters most, and it's the part most guides skip because it isn't glamorous. You don't need appliances to be smart. You need to know what they're doing.

An energy-monitoring smart plug costs roughly 20 to 25 dollars and turns any plugged-in appliance into a sensor. The Shelly Plug US Gen4 is a good example for this, because Shelly documents power metering plus local MQTT and scripting support rather than an app-only design. Power draw tells you the state, and the numbers below are illustrative, because every model differs. A dishwasher heating water might pull well over a kilowatt, one in standby a few watts, one off at the wall nothing. Build the automation on the power threshold and it works identically on a 2026 appliance and a 2009 one.

That single trick covers a lot:

1. Cycle-complete notifications for the washing machine, dishwasher, and dryer, with no vendor app involved.
2. Real cost-per-use for the air fryer, the kettle, and the second fridge in the garage.
3. Standby-draw detection. A coffee machine idling at 6W uses about 53 kWh a year, roughly 13 dollars at 25 cents per kWh, to do nothing.
4. Away-mode safety, cutting power to the kettle and the toaster when the house arms.

Before you buy, three things decide whether a plug is any use here:

• Energy monitoring, which is a separate feature from switching and the one you are actually paying for. Plenty of cheap plugs only switch.
• Local control, so the automation survives the manufacturer's cloud having a bad day.
• The load rating, because a kettle or an air fryer pulls far more than the lamp these plugs are usually sold for. The Shelly Plug US Gen4 is rated for 15 A (1800 W), so compare that to the appliance's nameplate before you plug a dishwasher or a kettle into it.

Not every cheap plug measures power, and the listing rarely makes it obvious. Energy monitoring is a separate feature from switching and it's the one you're paying for, so check the spec sheet before you order six of them. Pairing them is the same process as any other device, and if you're switching a hardwired appliance instead, smart switches are the right tool rather than a plug.

The sensor layer is the other half. A door sensor on the fridge catches the kid who leaves it open, a temperature sensor inside catches a failing compressor before the food does, and a leak sensor under the dishwasher catches the thing that costs real money. All of them are covered in more depth in the smart sensor roundup, and none of them care what brand your appliances are.

When Does a Fridge Door Sensor Beat a Smart Fridge?

Let me be blunt about the price gap. A connected fridge with a screen sits at the top of the market. Samsung lists its 29 cu. ft. Bespoke 4-Door Flex with a 32-inch Family Hub screen at 4,999 dollars. What a Home Assistant integration documents for refrigerators is mostly settings and state, and the screen and camera are not part of it. A Zigbee door contact and a temperature probe together cost about 30 dollars and give you both facts worth having: is it open, and is it cold.

The cold half is the one that protects food. The FDA says to keep a refrigerator at 40 degrees Fahrenheit (4 degrees Celsius) or below and a freezer at 0 degrees Fahrenheit. An automation that alerts when the probe stays above 5 degrees Celsius for an hour or two catches a failing compressor or a door that didn't latch before anything smells wrong. Set the delay long enough that a grocery unload doesn't trigger it.

Does that mean connected appliances are pointless? No, the good ones genuinely save time on the cycle-notification side. It means the screen isn't the useful part, and the useful part is cheap.

How Do You Build Automations That Don't Annoy Anyone?

A kitchen automation that fires wrongly gets unplugged within a week, so the bar for reliability here is higher than for lighting. Three patterns are worth building.

Cycle-complete with a delay. Power drops below the threshold for three continuous minutes, then notify. Without the delay, every pause in the wash cycle sends a false alarm. This is the single most common mistake people make with power-based detection.

Extractor fan on cooktop heat. A temperature sensor above the hob triggering the fan, rather than a motion sensor or a schedule. Motion in a kitchen means nothing, someone making tea sets it off.

Leak response, not leak notification. A leak sensor that only sends a phone alert is half a system. Pair it with a smart valve or, at minimum, cut power to the appliance. The value is in the reaction, not the message.

Once these grow past a handful of conditions, YAML gets unwieldy fast, and this is the point where a lot of people switch to Node-RED for the flows with real branching. For anything sensor-driven that you want to build yourself, ESPHome puts a custom temperature or current sensor in the kitchen for a few dollars in parts, with no cloud anywhere in the chain.

What About the Robot Vacuum and the Rest of the Room?

The kitchen floor is the hardest surface in the house, crumbs, spills, and the one spot where every robot vacuum earns or loses its keep. Local control matters more here than elsewhere too, because a kitchen clean is the one you want on a strict schedule. The models that run locally with Home Assistant are worth filtering for before you shop on suction numbers alone.

Everything in this guide assumes your Home Assistant instance is reachable and secured, which becomes more relevant the moment appliance clouds are in the picture. If you're exposing anything externally to get cycle notifications on your phone, read the secure setup notes first, and use the companion app rather than opening ports you don't need.

How Do You Check an Appliance Before You Buy It?

Five minutes of research saves a lot of regret here, and the checks are the same for every brand.

Search the Home Assistant integrations page for the manufacturer name first. If there's an official integration, read its documentation page and look at what entities it creates, that list is the real answer to "what will I be able to automate". An integration that only produces a single binary sensor is not the same as one that exposes program state and remaining time, even though both count as "supported" in a marketing bullet.

Then check the community forum for the specific model number, not the brand. Appliance makers change their cloud API between product generations more often than you'd expect, and a 2023 model that works beautifully tells you very little about the 2026 refresh sitting in the showroom. Threads where people report a working setup are worth more than any spec sheet.

Last, ask whether the appliance does anything useful with no network at all. If the answer is no, you've bought a subscription with a door on it. A sensible rule: the appliance has to be excellent at its actual job first, and any integration is a bonus that might evaporate with a firmware update or a pricing change, as SmartThings users are about to find out.

Which Common Mistakes Are Worth Avoiding?

These are the patterns that come up again and again when owners describe what went wrong.

Putting the coordinator too far from the kitchen. Appliance metal and a microwave oven can eat a marginal signal, and the failures look random, a device that works for a week then drops during dinner. If you're getting flaky behaviour in one room only, move the antenna before you replace any devices. The same applies to Thread, and Thread network troubleshooting covers what a weak border router looks like from the diagnostic side.

Automating things that were fine. A fridge that emails you is not an improvement over a fridge. Every automation is something that can misfire at 3am, so it needs to earn its place. Leak detection earns it. A notification when the freezer door has been open for four minutes earns it. A voice announcement every time someone opens the fridge does not, and it will be switched off within a fortnight by whoever lives with it.

Trusting a cloud integration for anything safety-related. Leak shutoff, gas alarms, and anything that cuts power should work with the internet unplugged. Test it that way once, properly, by pulling the WAN cable and seeing what still fires. A cloud integration can't pass that test by definition, which is worth knowing before you rely on one.

Buying the ecosystem before the sensors. People spend two thousand dollars on a connected appliance and then wonder why the automation side feels thin. Twenty dollars of measurement usually delivers more than that appliance's entire app.

What Would I Buy If I Started the Kitchen Today?

Four energy-monitoring plugs, a fridge door contact, a temperature probe, and a leak sensor under the dishwasher. Around 150 dollars at the prices above, all of it local, none of it dependent on a manufacturer staying interested.

Then, and only then, would I care whether the next appliance has an API, and I'd let Home Connect be the tiebreaker between two machines I already wanted for how they wash dishes. Buying an appliance for its integration is how people end up with a 2000-dollar oven and a dead app. Buying sensors for their integration costs 20 dollars and works on everything you already own.

The kitchen doesn't need to be smart. It needs to be measured. Those are different projects, and only one of them is expensive.