I've had 2x C7, C8, C8-Pro in both my main home and my airbnb for many years. Each hub is better than the last, and the software continues to improve. I've been a WebCoRE guy since the SmartThings days and was elated that it ran locally on HA.
I finally made the switch over to HomeAssistant in both houses, running on Pi 5s with NVMe SSDs. This weekend at the Airbnb I re-paired 27 zigbee devices, 26 zwave devices, 12 Lutron devices, and an Ecobee.
Two main drivers of the switch:
ActionTiles went away. I've been an ActionTiles user since the SmartThings days, and I was never able to rebuild exactly what I wanted with Hubitat's dashboard options. For example, sometimes I want a tile highlighted if it's in the opposite state (e.g. show up red when it was OFF since that's the warning state for a water valve). Maybe it was possible, didn't get in too deep.
The ability to configure everything in yaml. I hate yaml as much as the next person, but when everything is a file, gemini, codex, & Claude are really good at creating automations & dashboards. This was the killer one for me..... paste in entity names, describe the automation I want, bam, done.
By the way, it's somewhat fascinating that HA can talk to the Ecobee via the HomeKit protocol (even if you don't use HomeKit)....is HomeKit going to become the de-facto local API?
The HA ZWA-2 ZWave USB module is amazing (I mean, it should be with a foot long antenna). Zwave performance is so good.
This is key. AI is going to upend a lot of things like home automation. It might even be able to put lipstick on the Home Assistant pig (all the platforms have a little swine in them, even higher end stuff like Control4). I have no interest to learn YAML automations/scripts/templates/blueprints. Hubitat is a better environment to create complex automations for skilled non-programmers. Now AI lets one describe the outcome and it mostly does it. Unfortunately you still need to learn a little YAML to fix AI screw ups. Note also that the AIs are doing a pretty good job at Groovy, even Hubitat's constrained version.
I'm sticking with all of Hubitat, Control4, and Home Assistant for now. AI can make all of them work better. We'll see how things work out.
Hubitat can do that same thing now via the HomeKit Controller Integration. It's a work around on both platforms as there won't be too many HomeKit only devices in the future as Apple is on the Matter bandwagon now.
Yes, I’ll likely EOL my Rheem integration. The good part is… AI can take it from here. With the new capabilities I just added to the RS485/ESPHome integration, they were 100% written with Codex.
@bill.d I am a developer so I deal with JSON/YAML/etc all day long.
I’m definitely not “one of those guys that is now a HA convert.” I still dislike many, many things about HA. It’s just the most flexible for my use cases at this current moment. I will still recommend Hubitat to others for sure.
I consider Hubitat a legacy platform. Groovy in 2026?
I worked for a startup 2 years ago, and our Biz Dev person had a meeting with the Hubitat team. Based on what I heard, I do not see the company growing to much more than it is today.
How does the language matter, if AI can do the job? Seems more important that platform features are well covered by the API - in HE's case, Groovy or HTTP.
Why not? It is considered a niche language, but it certainly has it uses and it has not been abandoned.
Key Takeaways for Groovy in 2026:
Active Maintenance: Apache Groovy remains maintained, with version 4.0 and beyond focusing on modernizing the language.
Vital Niche: It is far from dead in ecosystems like Oracle EPM, which is introducing a new, stricter Groovy engine in Jan 2026, forcing updates to existing scripts.
Yeah, if you want to abandon Hubitat, just do so quietly and leave. Many of us have HA running with Hubitat, and we already know if we would want to slide over into the only HA world or not, so we don't need people telling us why they decided to do it.
Dashboards aside, custom apps and drivers are files, and AI can certainly be used to make custom automation apps and drivers with Hubitat. I don't need AI help with dashboards, they do what they do, and my focus is on automations with Groovy.
I thought it was clear that was a Gemini response, but I do usually include the header are put it in quotes, I wasn't trying to mislead where that came from.
What do you want modernized that would would change how Hubitat works? Functional does not mean you have to be on the latest and greatest thing.
Maybe. But check out Control4. Drivers programmed in lua. Administration Windows app that looks like Novell Netware from the turn of the century. Still the most capable AV automation controller. Programming language isn't holding anything back. Having said that, my guess is that it wouldn't take a small skilled team that understands home automaton much of a token budget to make something that is possibly better. And that might include a language choice that is different.
