Blinds, shades and garage doors: what works and what is still theoretical

Two device types, two completely different situations#

This page covers what looks like one category and is actually two.

Window Covering is a Matter 1.0 cluster. It covers open, close, go to a position, stop, and where supported, tilt. Google documents it as its "Blinds" ecosystem type, Amazon maps it onto Alexa.RangeController, and Apple lists "blinds and shades" among supported Matter accessory types. Several no-drill Matter over Thread blind products shipped through 2025 and 2026. If you buy a Matter blind today, it works.

Closure is new in Matter 1.5, released 20 November 2025, and it is the richer model: garage doors, gates, and coverings with real motion state and safety semantics. Ten months later there are zero certified closure products and no ecosystem exposing the device type. It is a specification, not a shelf.

Position reporting is less exact than the slider suggests#

The window covering cluster reports position as a percentage, and most controllers draw that as a slider. What sits behind the number varies enormously.

• Encoder-based motors count revolutions against a calibrated travel limit. These report position that broadly matches reality, and they hold calibration until something slips.
• Time-based motors run for a duration derived from a calibration pass. Position drifts as the motor ages, the battery sags or the fabric stiffens in cold weather. You recalibrate.
• Retrofit tube and bead-chain motors add slip. A chain skipping a bead once puts every subsequent position off by a fixed offset until you reset the limits.

The practical rule: treat reported position as approximate unless the vendor documents an encoder. Automations that say "set to 40%" will land somewhere near 40%. Automations that say "fully closed" are reliable, because a limit switch or a stall detection is doing the work rather than arithmetic.

Tilt is worse. Tilt is optional in the cluster, and controllers cannot reliably tell which features a given device implements. The result is a well-documented artifact: Apple Home can render a tilt slider that does nothing on a roller blind, because the controller assumes a capability the hardware does not have. That is not a bug in your blind. It is the controller guessing.

Battery versus wired motors#

This is the second real decision, and it is mostly about how often you want to be on a ladder.

Battery motors win on installation. A rechargeable pack, a solar strip on a south-facing window, no electrician. They cost you in two places: the recharge cycle, which for a heavy blackout shade in a sunny room can be every two to three months, and the radio. A battery blind motor is nearly always a sleepy device on Thread or Zigbee, so it will not relay for its neighbors, and a room full of them does nothing for your mesh. See designing a Thread mesh that holds up for why that matters when battery devices are the majority of your network.

Wired motors win on everything else: torque, speed, quiet operation, no maintenance cycle, and mains power means the device can act as a router and strengthen the mesh. They cost you an outlet near the window or an in-wall install, which in most US homes means a hole in the drywall or a very visible cable. Motorized roller shades with a low-voltage transformer in the header are the common professional answer.

Weight is the deciding variable people forget. A 96-inch blackout roller on a wide window is heavy, and a battery motor that handles a small kitchen shade will struggle, run slowly, and drain fast. Check the manufacturer's maximum shade width and weight, not just the advertised battery life.

Retrofit versus replacement#

Retrofit is the right first move for renters and for anyone testing whether they will actually use motorized shades. The renters guide treats this as one of the better reversible upgrades. If you are doing a whole floor of a house you own, replacement shades with wired motors will be less annoying for the next decade.

Safety: cords and garage doors#

Two genuine hazards live in this category, and neither is a software problem.

Corded window coverings are a strangulation risk to young children, which is why the industry has moved to cordless and motorized designs and why retrofit motors that leave a loose bead chain hanging are worse than the blind they replaced. If you retrofit a chain drive, secure the chain with the supplied tension device anchored to the wall, and do not leave slack within reach of a crib or a bed. A motorized cordless shade is a safety improvement over a corded one, which is a rare case where the smart version is also the safer version.

Garage doors are the highest-consequence device in a normal house. A residential opener is required to have entrapment protection: a photo-eye across the opening and an auto-reverse that stops and reverses on obstruction. Any remote-control scheme has to respect that.

The garage door standards gap, and the workaround that won#

Matter 1.5 defined closures explicitly including garage doors. Nothing has been certified. Kwikset's smart garage opener, announced in 2025, makes no Matter claim. In practice, garage doors are still outside the standard entirely.

The category is also the clearest example of a manufacturer closing an interface after the fact. Chamberlain deployed bot protection against third-party myQ access in autumn 2023, and the Home Assistant integration was removed on 6 November 2023. Nothing has reversed in nearly three years. Owners who had built automations lost them, and the hardware was unchanged.

What replaced it is instructive. ratgdo is a small local control board that wires into the opener, decodes Chamberlain's Security+ 2.0 and Security+ 1.0 protocols, reports real door position rather than a guess, and runs firmware for ESPHome, HomeKit or MQTT. No cloud, no account, no subscription, no vendor permission. It is not a standard and it is not certified, and it is the working answer, which says something uncomfortable about where the standards process is on this device class. The broader pattern is in when a smart home company shuts down and local versus cloud control.

If you would rather stay on supported hardware: a garage door controller that exposes a local API or HomeKit directly, wired to a proper tilt or magnetic position sensor, gets you most of the way. What you want from any of them is the same thing: the controller knows whether the door is actually open, not just whether it sent a command.

Buying rules#