VerifAI

Get on the grind. Then get online.

Comment

Inspiration

We've all had days where we felt an inability to "lock in" - to get off our phone and do that task that's been bugging us. This problem has been lamented time and time again, but not like this.

Now, with the versatility of artificial intelligence, those days are gone. With VerifAI, you have a live assistant to hold you accountable - instead of just removing these addictive apps from your life, we _ actively replace them _ with things that move you forward.

What it does

All the user has to do is select a task and select the apps they want to restrict. Whether that's cleaning your room or doing the dishes, our AI will contextualize your situation and generate deliverables for you to achieve. When you complete your task, it will unlock your favorite entertainment apps.

How we built it

VerifAI is created with the Swift language (for iOS/iPadOS), on Apple's development platform Xcode. Additionally, we used the Grok API to analyze user-submitted photographs and determine if they met their goals.

Challenges we ran into

As previously mentioned, in order to create an iOS app, you need to use Xcode. Unfortunately, due to hardware limitations, one of our team members had to use a legacy version of Xcode, and we had to address significant compatibility issues when syncing with code developed on modern versions.

Additionally, we used a notoriously difficult framework called ScreenTimeAPI (and its subsidiaries). This is quite difficult to customize and communicate with, challenging us by forcing us to be very direct and clear on exactly what information we needed to pass and where.

Accomplishments that we're proud of

We worked special Grok frameworks to typecast outputs into XML, and improve predictability of GenAI outputs. None of us had any experience with xAI, and we became much more knowledgeable on the process of collaborating with AI to accomplish our goals. That's a skill that will become more and more relevant with time, so this experience was quite valuable.

What we learned

It was our team's first experience using Swift at a hackathon, and we gained a lot of experience working with the tools that real software engineers use day-to-day. We anticipate that this will leave us more prepared for our development projects long into the future.

What's next for VerifAI

We could extend VerifAI to handle multi-step iterations in tasks, with xAI determining how far along you are in completion. We could also integrate other AI models to enhance flexibility for the user.

Built With

Share this project:

Updates