Home Assistant 2026.11 Overhauls Device and Entity Naming

Home Assistant 2026.11 is rolling out with one of the most far-reaching changes the open-source smart home platform has made in years: a complete overhaul of how devices and entities are named. The new system, which arrives with the November release, replaces more than a decade of layered naming rules with a single principle — name each thing once, for what it actually is, and let Home Assistant add the context around it.

The change was originally planned for the 2026.4 release in April, but the Home Assistant team pulled it at the last minute after beta testers pushed back. Seven months later, it is back — reshaped by that feedback and bundled with a new concept called child devices. For the millions of households running Home Assistant, the update promises tidier dashboards, simpler renaming, and an end to the naming workarounds that long-time users know all too well.

Why naming needed an overhaul

Home Assistant’s naming problem grew slowly, one sensible decision at a time. In the platform’s early days, there were only entities — individual sensors, switches, and measurements configured in YAML files — so an entity’s name had to carry all of its context. To tell two power sensors apart, users wrote names like “Kitchen coffee machine power.” The community even produced whole naming strategy guides to keep things consistent.

From 2018 onward, Home Assistant added structure around entities: devices to group the entities belonging to one physical object, areas to describe rooms, and later floors and labels. Each addition made sense on its own, but together they left rules that did not always agree. Names repeated themselves across layers, and moving a device to another room could leave every one of its entity names wrong. Integrations began naming entities on their own — “Power” instead of “Kitchen coffee machine power” — with Home Assistant adding the device name in front. One gap remained: renaming an entity stripped its device name, so users often had to rename every entity by hand to stay consistent.

“Each of these steps made sense at the time,” wrote Home Assistant’s Artur Pragacz in the announcement, “but together they left us with rules that didn’t always agree. We’ve been untangling them for several years, and this release is the next step.”

How the new naming works

Under the new system, a consistent set of rules applies everywhere in Home Assistant, from the dashboard to the settings pages. Each part of a name is stored once: the floor is “Ground floor,” the area is “Kitchen,” the device is “Coffee machine,” and its entities are “Power” and “Energy.” No name repeats another. Home Assistant then shows context only where it is needed.

On the Kitchen’s area page, where the room is already clear, the device is listed simply as “Coffee machine,” not “Kitchen coffee machine.” On the device page, where the header already names the device and its area, the entities appear as just “Power” and “Energy.” In an entity picker listing every power sensor in the house, each row also shows the device and area so sensors can be told apart. And on a dashboard card, users can choose which parts to show — entity, device, parent device, area, or floor — in the order they prefer.

Some entities need no name of their own at all. The switch on a smart plug is what turns the coffee machine on and off, so it will be shown with the device name alone: “Coffee machine.” Home Assistant calls this the main entity of the device, and users can turn it on for any entity from its settings. This is why renaming the device, rather than the switch, now gives most people the result they want.

Child devices for multi-part hardware

The biggest structural addition is child devices, designed for hardware made up of several parts. A power strip with four sockets becomes one device with four child devices, one per socket. Each socket can have its own name, such as “Kettle,” and its own area. Move a socket’s lamp into the dining room and it will simply show as “Lamp” there, without the power strip’s name attached.

At launch, grouped child device support is available on Z-Wave, MQTT, and ESPHome, along with some Tuya devices and other integrations. Support for Matter and ZHA is still in progress, with more integrations to follow in future releases.

New entity IDs are built from the same parts. Since the 2026.8 release, entity IDs have included the area, and users can choose which parts go into them — so two “TV” devices in different rooms get clear IDs without the old “_2” suffix.

What existing users need to know

The most important reassurance for existing users is that nothing breaks. Entity IDs do not change, so automations, scripts, scenes, and dashboards — which all refer to entities by ID — will keep working exactly as before. Only the names displayed on screen may change.

Existing names are migrated automatically on update. Most should carry over as expected, but some will look different. If an entity name already starts with its device name, such as “Coffee machine power usage,” the device part will be removed, leaving the entity named “Power usage” and displayed as “Coffee machine Power usage” — much as before. If an entity name does not start with its device name, nothing changes except that Home Assistant will add the right context where needed, so a switch renamed “Coffee machine” on a device called “Smart plug 3” will show as “Smart plug 3 Coffee machine.” That second case is quick to fix: rename the device to “Coffee machine” and turn on “Main entity of the device” for the switch.

For entities exposed to voice assistants, the old name is also added as an alias, so existing voice commands keep working. The friendly name — the combined device-plus-entity name that older cards and templates still use — will be kept for backward compatibility for now, but Home Assistant plans to deprecate and eventually remove it. Template users who display friendly names in notifications should switch to the entity_name(), device_name(), and area_name() functions instead.

Some custom cards may show different names than the built-in cards, since they may still read the friendly name. Setting a name on the card fixes it, or users can wait for the card’s author to adopt the new naming. Names already set manually on a card stay exactly as they are.

Part of a bigger 2026.11 release

The naming overhaul is the headline change in a release that also brings Matter cameras into the Home Assistant frontend, with a labs preview of two-way audio, and folds the Home Assistant Community Store (HACS) into the core interface as a built-in Marketplace. The naming work, though, is the change that touches every installation — even if for many users it will barely be noticeable.

Home Assistant is clear that the change will not be perfect for every setup. Naming, as the team puts it, is personal. But after seven months of rework driven by beta feedback, the goal is that users will no longer need a naming scheme to keep things tidy. The defaults should be right, and if you do want to rename something, you will only need to do it once, in one place — because no name holds a copy of another.

Over the coming releases, the team plans to move more integrations to child devices and help custom card authors adopt the new naming. In the meantime, users updating to 2026.11 should check their dashboards after the upgrade, tidy up any entity names that look odd, and consider it a good moment to strip area names out of entity and device names — Home Assistant will display the right context where it is needed.