What it does

Snow is an AI-powered day planner that turns natural language into optimized schedules. Students can upload their syllabus or simply chat about their day (“gym, study for midterm, dinner with friends”), and Snow intelligently builds a balanced schedule that respects class times, deadlines, and personal commitments. Users can invite friends to events, compare availability, and avoid double-booking, making planning both social and stress-free.

How we built it

We built Snow using a React + TypeScript frontend with Tailwind CSS and Framer Motion to create a smooth, playful user experience. We relied heavily on Lovable to rapidly iterate on both UI and logic, continuously testing, prompting, and refining the AI to better reflect how students actually plan their days. The backend is powered by Lovable Cloud, using Supabase for authentication, real-time synchronization, and a PostgreSQL database. AI scheduling and reasoning are handled through edge functions with Gemini, allowing Snow to translate vague, human requests into structured schedules. The snowfall animation? Pure CSS magic, lots of vibe coding

Challenges we ran into

One of our biggest challenges was security and privacy—ensuring that users’ calendars and availability were never exposed publicly, even when coordinating with friends. Designing correct row-level security (RLS) policies while still enabling real-time collaboration required careful testing and iteration. We also had to handle real-time calendar sync across multiple users, gracefully resolve invitation conflicts, and teach the AI to understand ambiguous requests like “squeeze in gym somewhere” without violating class schedules or deadlines.

Accomplishments we’re proud of

A seamless friend invitation system with automatic conflict detection Automatic syllabus parsing that extracts both class times and assignment deadlines An AI planner that feels genuinely helpful, not rigid And, of course, the confetti burst when you accept an invite, small detail, big dopamine

What we learned

Social features are significantly harder than solo features—real-time state syncing across users demands careful architecture and security decisions. We also learned that heavy iteration with AI tools matters: repeatedly prompting, testing edge cases, and refining behavior made the planner feel far more human. Finally, good vibes, animations, and polish go a long way with students.

What’s next for Snow

A mobile app, Google Calendar integration, group study session planning, and an AI that adapts to user behavior over time (“you always skip morning gym—want to try evenings?”). Our goal is for Snow to feel less like a planner and more like a personal academic assistant.

Built With

• css
• typescript