New Google AI Studio integrations with database and auth using gemini 3.5

This is a submission for the Google I/O Writing Challenge

Google AI Studio now supports native Kotlin to vibe code Android apps. With Google AI studio this is now easy to make and with Firebase services to make into a fullstack app which can be downloadable or published with the model 3.5 gemini is something i find intriguing.

1) So I used the new model to build an alarm app that runs in interval in 10mins and it cam uo with designs which is this I choosed you can customize the design set up and it took me up to 282 seconds for the alarm to be completed

This is the preview of it when it was officially time and I set up differnt times and another imperssive thing was the firebase additon i could litterlay turn this up to a fullstack application use it and vibecode my way to have a little android appliaction

This is the preview of the code of the native kotlin from the new kotlin which is now faster and smarter compared to earlier version of gemini which I give credit to

der how it will fare with bery complex applications in the near future

An impressive thing it did was this brought options to add snooze functionality and multiple alarms and asedk for permissions which was perfectly simulating an andrioid app while working on just one app which will give somoene with little or no tech skills in kotlin to understand what is pertectly going on

using api key and will switch back when back in quota is something I find impressive hold things in memory for you when out of credit and get back to work as expected which is awesome for me seems everyone will become a vibecoder these modern days and just understand the architecture of how things work

Using this model its just like writing on a book understanding the basics and just using the knowledge to bid simple apps, I won

2) Next up the migration agents the basic file vibecoded in vscode in reactive native with the new model was able to migreate this into Kotlin native app which is impressive based on what model you want to use

I focused mainly on the Google Developer Keynote and I cant wait to see what it has install for the next model and how far AI research will go

Thanks for reading my little contribution the Google AI challenge