This is using the (beta) Shelly wifi integration available in the latest firmware.
I’m seeing incorrect child/entity mapping with a Shelly 1 Gen4 in Hubitat’s native Shelly integration.
Expected behavior
Shelly 1 Gen4 has:
- a physical input (
SW) for sensing external state - a dry-contact relay output (
I/O) that can be controlled separately - in Shelly, the two may be configured to be in sync, or detached, or following other policies
Actual behavior
Hubitat appears to conflate input and output state.
- In Shelly’s own UI, the device behaves correctly:
- input follows the external switch state
- output follows or does not follow input depending on Shelly settings
- In Home Assistant, the same device is modeled correctly and input/output are distinct
- In Hubitat native Shelly integration, the exposed child devices do seem to capture the intended behavior in the Shelly device (one references "input", the other "switch"), but the "input" device never shows a state (which is meant to match the external switch), while the "output" device seems to show the state of the external switch - but also the output state of the Shelly if I change it manually from the Shelly interface... Difficult to described, but it's quite messed up.
I also reproduced the same problem on a second Hubitat hub, which suggests this is not a one-off hub state issue.
Reproduction
- Add Shelly 1 Gen4 to Hubitat using native Shelly integration
- Wire external relay/contact to the Shelly input
- Observe child devices/entities in Hubitat
- Compare behavior against Shelly local UI
- Put Shelly in Detached mode and retest
Result
Hubitat still does not preserve correct input/output separation.
Additional note
As a cross-check, I added the same Shelly 1 Gen4 to Home Assistant and HA reflects the device correctly, with separate input and relay behavior. That strongly suggests the issue is in Hubitat’s Shelly integration layer rather than in the device itself.
