the goal is to test AI in snap
AI is too general
???
Variables in Snap are persistent by default. Just save the project after training.
has model stored.
why train in snap at that point
SciSnap has a native.
So end-to-end flow inside Snap seems to be possible
And to deny the false statement
does it save the data so you can have multiple people train it and it remember all of them
No...
Load the shared 
project.
Then save a trained model.
Share it, so someone else can further train.
Share the link on the forum.
Repeat.
But... if you already have shared storage save/load the training data
ahh i see. still dont get why you wld train inside snap!
You could create your own cloud server and store it that way, but you would have to protect the server from receiving false data
I think i might just make my own cloud server but it will be permission based and you must get permission to use the database first. I might change it in the future.
make a password that embedded in the request idk
why is this turning into a MQTT topic
...it's not?
it almost turned into one
mqtt ≠ cloud storage
ugh, ok
it can be
The core purpose of MQTT is to transport messages between multiple devices reliably.
but it can also be used to store data