Special attention is paid to security and camera management all the way to implementing Vision AI. In my tests Zumungo announces "UPS truck approaching the drive way. A person carrying a package walks towards the door". This announcement is processed by an event and results in "UPS delivery" on the GH speakers around the house.
Zumungo has it's own event engine that is easy to use by both the user as well as AI Agent. User can tell AI agent " Create an event to open the garage door and turn on the porch light, upon my arrival home on foot or by car tracked by zumungo geo fencing".
Option exist for users to design their own panels to display on tablets on the wall or desktop computers for even easier management of the house by family members, guests, kids etc. For example if you rent your ADU on Airbnb than you can create dedicated panel on a tablet for guests to use that controls their rooms only and has no access to the main house.
Zumungo has it's own mobile app available at iOS store. For Android devices it is downloadable from Zumungo server itself.
As far as AI goes, Zumungo offers simple interface in settings for user to pick and choose online AI agent from the major 3 ( Claude, Gemini or Open AI ) or pick a local agent from a list of various agents. Available local agents include Llama, Mistral, Qwen, DeepSeek, Phi and others. Zumungo always updates the agents to the best and latest and always vets them for licensing and legal compliance. Zumungo also advises user what kind of hardware is needed for particular agent to run typically 16GB or more of GPU memory.
Trying Zumngo is easy. One line on any Linux machine or ready made RPi image - download-flash and go. See "get started" at Zumungo web site.
Support issues come after trying and this is taken care of as well at Zumungo Forums or just email support@zumungo.com