Does a home automation tag need the internet?

The tag is never the problem

Reading a tag is an exchange over a couple of centimetres between a phone and an object. No network is involved at any point.

So a tag that appears to fail during an outage has been read perfectly. What failed is everything after it.

That distinction matters because it points the investigation at the home system rather than at the wall.

The request that leaves the house

Many home systems route a command through a service on the internet even when the phone and the device are in the same room.

The phone asks the service, the service asks a hub or the device, and the lamp responds. A round trip out of the building and back for a distance of three metres.

When the connection drops, that path breaks and the action fails, which surprises people who reasonably assumed a local action was local.

Some systems work locally and fall back to the service, some work only through the service, and the difference is rarely advertised.

Finding out which you have

Disconnect the internet at the router, leaving the local network running, and try the action from the application.

If it works, your system acts locally. If it does not, every automation you build depends on your connection.

Test each kind of device separately, since a system can be local for lights and remote for something else.

Ten minutes of this tells you more about your home system than any specification sheet.

Designing for the outage

Keep the physical controls working. A light that can only be switched by a phone is a light that fails with your connection.

Prefer systems and configurations that act locally where the choice exists, particularly for anything about doors, heating or safety.

Do not put an outage critical action behind a tag, since the tag is the reliable part and the chain behind it is not.

And tell the household what stops working when the internet does, so an outage produces an explanation rather than a search for a broken tag.