About Alex Rivera and the House Behind This Blog

Why this site exists

My first hub was a Samsung SmartThings, and it reached my desk about ten years ago because of work rather than a shopping trip. I spent that decade automating a flat the way most people do it: one box at a time. Philips Hue for light. An eQ-3 MAX! Cube for the heating. Then Tuya and Xiaomi gear, bought because it was cheap and it did the one thing I needed that week.

Ten years of that teaches you something no product page will. Every ecosystem you add is another app, another cloud account, another thing that breaks on its own schedule. Devices stop talking to each other when a vendor rewrites an API. A hub goes end of life and takes your automations with it. By the end I had a flat full of things that worked, and no system at all.

So when I started building a house, I decided it would be properly smart rather than smart-ish. Not a shelf of gadgets bolted on after the painters left. Smart from the wiring outwards: cable runs planned before the walls closed, deep back boxes with neutrals where the switches would go, a small server rack in the utility room, and blind motors specified while the window openings were still holes in a wall.

That build is two years old now, and the list is long: lighting, air conditioning, a heat pump feeding underfloor heating, solar, blinds, the gate and the garage door, alarm, cameras, connected appliances, garden irrigation, a mowing robot and a vacuum. Each one is cheap to decide on paper and expensive to reverse once the plaster is on.

Almost nothing I read while planning that was written by somebody who had actually done it. Plenty of unboxing videos. Plenty of "top 10 smart plugs". Very little about the choices you only get to make once. So I started writing down mine, as I make them, with the mistakes left in.

Every guide on this site traces back to a decision I had to make on site, or a device I put into the house and lived with.

The stack this house runs on

I picked three systems and committed. That is the single most useful thing I did, and I'd argue most smart home advice gets it backwards by telling you to shop first and integrate later. The full reasoning, including why I turned down a single-vendor package, is in why I chose Home Assistant:

• Home Assistant is the brain. Every automation, every dashboard, all of the logic. It runs locally on hardware in my own utility room, so the lights still work when the fibre goes down.
• Google Home is the face. My family isn't going to open a YAML file, and they shouldn't have to. Voice control, the phone widgets, the routines everybody uses without thinking about them.
• Somfy moves everything mechanical. Twelve blinds so far, on io-homecontrol motors driven through TaHoma. Pulling TaHoma into Home Assistant, so shading follows the sun and the indoor temperature instead of a wall remote nobody can find, is the next job on the list rather than a finished one. The motors went in while the window openings were still holes in a wall, which is the part you can't retrofit.

Everything else is bought on merit, category by category, and none of it is from the same company: iRobot for vacuuming, Gardena for irrigation and mowing, Fujitsu and Gree for heating and cooling, Satel for the alarm, Reolink for the cameras, Huawei on the roof. Lighting is the deliberate exception, and it is a decision rather than a finished job: the wiring is in and waiting for unbranded Zigbee controllers that pair straight to a USB dongle on the Home Assistant machine, no bridge and no app, because ceiling light has no category leader worth its own bridge. The controllers are the next thing I buy. Nobody makes the best product in every category, so a single-vendor house means settling for fourth-best in most rooms. Buying six leaders instead means six apps, which is exactly the problem Home Assistant solves: one dashboard, and systems that can finally act on each other's data instead of ignoring each other.

The shortlist a device still has to pass: local control, survives a router reboot, and shows up cleanly in both Home Assistant and Google Home. Cloud-only devices that stop working when a vendor loses interest don't get wired in. Would you put one behind a wall?

What you'll find here

Roughly three kinds of writing, and they all come from the same build:

• Build guides covering the decisions that happen once: where the cable goes, what the electrician needs to know, how many circuits a smart lighting plan actually wants, how to wire blind motors before the windows arrive.
• Setup and integration tutorials for Home Assistant, Google Home, and Somfy, plus the protocols underneath them, mostly Zigbee, Z-Wave, Thread, and Matter. Configuration examples come from my own installation, not from documentation.
• Device reviews and comparisons, written to answer one question: does this thing belong in a house built around this stack, or does it fight it?

Background

I've been a professional software engineer for over 20 years, across backend systems and embedded platforms. Home automation started as a side interest ten years ago, ran through four ecosystems in one flat, and turned into the reason I planned an entire house around a server rack. The engineering habits carry over: I want local control, version numbers, and a system I can debug at 11pm without phoning support.

That background is also why I'm blunt about failure. Automations break. Firmware updates undo things that worked. When something in this house falls over, it goes on the site with what it took to fix it, because that's the part nobody publishes.

How I test

Anything I test directly goes into the house and stays there. It spends at least two weeks in daily use before I write a verdict, and I check every claimed integration myself instead of trusting the box: pairing in Home Assistant, exposure to Google Home, and where it applies, whether it plays along with the Somfy setup. What I don't do is record firmware version numbers, so you won't see one quoted here. Inventing one afterwards would make a verdict look more precise than it is, and "this thing loses its connection" is more use to you than naming a build you'll never be shipped.

I can't buy everything, and this site's catalogue is far bigger than my house. Where a review is research-based rather than hands-on, it says so and leans on manufacturer specifications and verified owner feedback. The full method sits on the how we test page, and the sourcing and corrections rules are on the editorial guidelines page.

Contact

Have a question, spotted an error, or want to suggest a product to review? Reach out on X (Twitter) or Bluesky (smarthomeguideblog.bsky.social - link in footer), or visit the Contact page (link in footer above).