Ezlo Softhub and CCTV+AI capabilities new versions out!

[curiousB]

@cw-kid

Although I was expecting an EzloPi sensor device to just appear as a new device on my existing Ezlo Plus controller. This is not the case, as the EzloPi’s do act like a “controller” in their own right. Which makes having sensors on multiple EzloPi “controllers” somewhat disjointed.

Yes,

I have experimented with the EzloPi software and agree it offers great promise. In fairness I haven’t tried the latest which I was told is much improved over the version I played with months ago. So maybe many of my concerns were resolved.

I do see this notion of EzloPi being a controller or peer to a EzloPlus as a liability. I feel there has to be a master/slave relationship with the controller and devices. I view EzloPi much like you do, a vehicle to create you own device. Something unusual and not likely to be commercialized by a large vendor. Currently you cannot include device status of an EzloPi in a local meshbot on a peer LAN to EzloPlus. You have to go cloud based. That is a serious limitation in my mind. This forum is littered with all the reasons meshbots should run local wherever possible.

This is why I wrote my own General purpose IO module in Arduino IDE running on ESP32 (A DIY General Purpose WiFi (IP) Based IO Module for Ezlo) . It is queried by HTTP GETS and I use the IP Template functionality of EzloPlus to manage that “slave” device. So this way I could have local meshbots control my custom IO Module. Unfortunately the IP Template area is another 70% done area in Ezlo. You can only interrogate certain sensor types and its missing a polling function so you can have background audits to sync state when local control changes the IP device’s state. ZWave handles this more elegantly with “Instant Status” functionality. I can live with Polling but the homerun is to design an “Instant Status” like API into the IP Template construct so local device state changes can be reported back to the controller on an unsolicited basis.

Anyway I realize finite resources to work on Ezlo functionality. I still stand by my longstanding argument that core functionality has to come first and be fully hardened and robust. When I see so much effort on new video logging, AI like image triggering, and so on I wonder if the latest feature pursuit comes at the expense of basic functionality hardening and completion.

Just my 2 cents.