Stratus

💡 Inspiration

We’ve all experienced the "chaotic morning routine" paradox: you check your phone, see a sunny forecast, and head to campus—only to be caught in a freezing wind chill or an unexpected downpour by your 2 PM lab across campus.

Broad city-wide weather reports fail because a student's life isn't lived "in the city"; it’s lived in specific buildings, at specific times, with distinct outdoor transits. Stratus was born to bridge the gap between knowing the weather and actually preparing for it. We wanted to eliminate the cognitive load of a morning routine by building an intelligent partner that understands your day as well as you do.

🚀 What it does

Stratus is a high-end personal utility that synthesizes your chaotic schedule into a clear, layer-by-layer attire strategy.

1. Multimodal Input: You upload a schedule—be it a high-res PDF, a blurry screenshot of a monitor, or a raw text file.
2. AI Synthesis: Our intelligence engine parses the mess to extract exact class times and campus locations.
3. Hyper-Local Logic: It fetches hourly, campus-specific forecasts for your university.
4. Master Strategy: Instead of just a temperature reading, you get a contextual recommendation: "Wear a windbreaker for the transit to the Engineering building, but keep a light layer for the over-chilled library."

🛠️ How we built it

We prioritized Rich Aesthetics and Agentic Intelligence. The stack was chosen for maximum reactivity:

• Intelligence: Google Gemini 2.5 Flash Lite serves as the multimodal brain, handling unstructured data parsing with surgical precision.
• Frontend: Built on the bleeding edge with Next.js 15 and React 19, using Framer Motion for a "glassmorphism" UI that feels alive.
• Cloud DNA: Supabase handles lightning-fast data sync, while Auth0 provides enterprise-grade security for user profiles.
• Precision Data: We integrated the Tomorrow.io API specifically for its micro-local resolution, which is essential for true campus-level weather matching.

🧠 Challenges we ran into

The most significant hurdle was the "Unstructured Data" problem. University schedules are notoriously inconsistent across institutions. Teaching an AI to reliably distinguish between a course name and a room number in a table-less text file required intense prompt engineering and iterative testing.

Another major challenge was the Logic of Transit. We didn't just want to show the weather at the start of a class; we had to build an engine that looks at the vulnerability windows between classes—the transitions where a student is actually exposed to the elements.

💪 Accomplishments that we're proud of

We are incredibly proud of the Generative Styling Engine. It doesn't just check a lookup table; it reasons. It understands transitions between indoors and outdoors, wind-chill volatility, and specific campus climates. Seeing the app successfully turn a messy, incomprehensible PDF into a structured, helpful styling strategy was our "Eureka" moment.

📚 What we learned

This project taught us that the future of personal software isn't about giving users more information—it’s about providing Synthesis. We learned how to manage complex, multi-source data flows (AI, Weather, DB, Auth) while keeping the user interface humane, approachable, and visually stunning.

🗺️ What's next for Stratus

We envision Stratus as a daily ritual. Our upcoming roadmap includes:

• Social Style History: Allowing users to track and share their "best fit" days.
• Multimodal Wardrobe Integration: Letting users upload photos of their own clothes for personalized "virtual closet" matching.
• Native Mobility: Moving beyond the browser with a native mobile wrapper for real-time campus push notifications.

Crafted on the bleeding edge of the Google Cloud & Vercel Ecosystem.

Built With

• auth0
• gemini
• next.js
• react
• supabase
• tailwind
• tomorrow.io
• typescript