The hardware for the pet marvel 2-in-1 is actually pretty good, but the software in both the app and the firmware are in need of lots of work. The AI recognition is barely functional, you can’t seem to correct wrong identifications (and if you can, it’s unclear how presumably due to translation issues). There’s also no per-pet tracking of water or food intake, so all the meals come out at once and it is basically a race between the animals when I was hoping it would see a pet and dispense the meal they are supposed to be getting.
Taking your pet’s picture for the AI recognition is also nearly impossible. Ever tried to get a cat to sit perfectly still so you could take a head-on photo of their face within a wireframe? Come on. We all know AI is capable of being trained on multiple images, and even simple hardware is capable of running some of the smaller open source models nowadays.
My suggestion is to open source the firmware and the app and let the community do what they do best; improve the software to make what is a pretty good piece of hardware into a must-have for multi-cat households.
Show less
squizzeak
Liudward