StudyPilot

Study Plan Page
Events Page
Tasks Page
Calendar View
Wellness Page
Syllabus Upload
Dashboard Page with user signed in with their canvas (srinivas)

Inspiration

Us college students juggle classes, assignments, exams, clubs, jobs, and personal life. Yet every student still has to manually check Canvas, plan tasks, and figure out what to study each day. We noticed that most students don’t struggle because they lack ability, they struggle because they lack clarity. We wanted to create something that acts like a personal academic co-pilot: a tool that pulls your Canvas data, analyzes it with AI, and tells you exactly what you should focus on today, not tomorrow, not next week, but today.

That idea became StudyPilot.

What it does

StudyPilot is an AI-powered academic planner that automatically: 1) Imports your assignments & events from your Canvas calendar feed 2) Organizes tasks by priority, urgency, and due dates 3) Generates a daily study plan and weekly schedule 4) Creates stats like study hours, exam count, workload balance, etc. 5) Provides a clean dashboard with: - Today’s schedule - High/medium/low priority tasks - Weekly study blocks - Upcoming exams - Quick stats

In short: StudyPilot turns your Canvas feed into a personalized academic game plan.

How we built it

We built StudyPilot using: Backend • Node.js • OpenAI API (GPT-4.1) • Canvas API • Custom prompt engineering for academic planning • Routes to generate study plans and priority scoring

Frontend • Lovable AI (for rapid UI generation) • React components for dashboard, calendar view, task manager • Dynamic pages for Today, This Week, Tasks, and Study Plan generation

AI Logic • We engineered a highly structured prompt that forces the model to output: • Enhanced calendar • high/medium/low priority groups • Weekly calendar • Today schedule • Weekly study plan • Ranked deadlines • Academic analytics block • The backend automatically cleans and validates this JSON.

Challenges we ran into

1) Parsing Canvas ICS Feeds: Canvas feeds aren’t documented well, inconsistent, and contain mixed event types. We had to write a custom parser to pull course codes, assignment titles, exam indicators, and URLs.

2) Forcing GPT to output reliable JSON: Models love to hallucinate or break JSON. We solved this by building a strict format with dozens of forced keys and custom validation.

3) Time constraints: Building an end-to-end AI planner within a 4 hour hackathon was intense. We had to optimize workflow, split tasks efficiently, and rapidly iterate.

4) Integrating with Lovable UI: Making sure the backend JSON plugged perfectly into the UI required quick coordination and several iterations.

Accomplishments that we're proud of

1) Fully functional AI academic planner built in a single hackathon session 2) Clean, professional UI created with Lovable 3) Successful Canvas ICS -> JSON -> AI -> dashboard pipeline 3) Structured AI output that reliably generates a real study plan 4) A tool that students would actually want to use to stay on top of school

What we learned

1) How to work with real Canvas calendar data 2) How to engineer complex, multi-section prompts for structured GPT output 3) How to integrate AI into a real-world student workflow 4) How to build a polished UI extremely fast using Lovable 5) How to collaborate under time pressure and iterate rapidly

What's next for StudyPilot

We want to expand StudyPilot into a full academic assistant: 1) Direct Canvas OAuth integration 2) Auto-parse syllabi PDFs and extract assignments 3) Smart time-blocking with Google Calendar sync 4) Personalized learning recommendations 5) Gamification 6) Long-term learning paths for each course