ShortBreak | Devpost

Inspiration

We’ve all been there: you open your phone for "one minute" to check a notification, and suddenly an hour has vanished through the unstoppable addiction known as "doomscrolling". We realized that while platforms like Instagram and YouTube are designed to be addictive, short form content aren't all that bad and have incredible potential for micro-learning. We built ShortBreak to flip the script: turning mindless scrolling into mindful learning and helping users regain the self-control they lose to the "doomscroll."

What it does

ShortBreak is a dual-platform solution designed to stop/prevent your scrolling habits and redirect them toward personal growth. It focuses on fixing the root cause of your addiction, by having you feel a sense of responsibility over your screen time.

The iOS App: A native SwiftUI application that acts as the scrollling blocker. Using iOS Shortcuts and Automations, the app intercepts your "autopilot" social media opens (like Instagram) and forces you to make sure you don't overspend you screen time. Every 15min, you are allowed to go claim a random amount of screen time for the day, ranging from 1min to 30min (the higher the number is, the lower the chance of getting it is). However, once it's gone, your access to Instagram is gone, so spend your time wisely. It keeps track of your total daily screen time and how much remaining time is left.
The Web Workspace: A Next.js-based platform that transforms YouTube Shorts from "distractions" into "lessons." It uses AI to generate structured learning outlines for every video, allowing you to track your history and turn a 15-second clip into a tangible piece of knowledge. It uses tactical prompt engineering to have you learn topic by topic based on either your personal prompt, or through uploading your course outline.

How we built it

Web: We built the web application using Next.js and Tailwind CSS. For the backend, we utilized Convex for real-time data management. We integrated the YouTube API to fetch content and the Gemini API to perform high-speed analysis of video content to generate educational outlines.
Mobile: The iOS app was built natively with SwiftUI and Swift 5.0. We used Core Data to manage user session statistics locally and integrated with iOS Shortcuts/Automations to create the "interception" logic that triggers when a user opens a distracting app. Uses a SQLite database to store user data.

Challenges we ran into

The original plan was having everything be coded through a single mobile app. However, the youtube API responses returned data that wasn't possible to properly use on a mobile environment. This costed us over 2h of our time and we even considered giving up on the project. Another big issue is having to deal with an infinite loop iOS automation. We had the app appear whenever Instagram is being opened. However, when we exit the app to continue on IG, it reopens the app, causing this infinite loop. We had to go with workarounds, using creating custom shortcuts and having the user do extra setting up steps while keeping things as user friendly as possible.

Accomplishments that we're proud of

We are very proud about the UI and UX of the iOS app. It has a very warm and friendly vibe, with animations that just seems right. Similarly, the theme is well incorporated accross web and mobile, making it look like a very well built project. This is something I will be keeping on my phone for the future.

What we learned

The idea that interrupting a habit loop is more effective than trying to avoid it entirely. If we were to simply delete insta, people will often go to another source of short form content. Instead, restricting it without entirely blocking it will slowly reduce his time spent scrolling and erase his addiction. Technically, we deepened our knowledge of the Next.js App Router, mastered the reactivity of Convex, and learned how to leverage the Gemini API for structured data extraction rather than just simple chat.