Studentify - AI Study Planner

An AI-powered academic assistant that reads Canvas data, syllabus content, and student schedules to create personalized daily study plans. Every morning, it sends students a clear study plan through Telegram, and during breaks between classes, it recommends the best lesson to review and gives a short quiz.

Problem Statement

Students have:

• Deadlines scattered across Canvas
• Syllabus info hidden in PDFs
• Small breaks between classes that get wasted
• No clear idea what to study first
• No personalized daily plan

Studentify solves this by automatically creating smart study recommendations based on:

• Canvas deadlines & grades
• Syllabus exam/quiz dates
• Class schedule breaks
• Weak subject detection
• Time-of-day availability

---

Quick Start

Prerequisites

• Python 3.9+
• Node.js 16+
• Canvas API token (from your school)
• Telegram Bot token
• Supabase account (free tier works)

Clone & Setup

git clone <your-repo>
cd studentify
cp .env.example .env
# Fill in your credentials in .env

Run All Services

# Terminal 1: Backend
cd backend
pip install -r requirements.txt
python main.py

# Terminal 2: Frontend (in new terminal)
cd frontend
npm install
npm start

# Terminal 3: Bot (in new terminal)
cd bot
python telegram_bot.py

Backend runs on http://localhost:8000
Frontend runs on http://localhost:3000

---

Project Structure

studentify/
├── backend/
│   ├── main.py                 # FastAPI entry point
│   ├── canvas_integration.py   # Canvas API wrapper
│   ├── database.py             # Supabase client
│   ├── routes.py               # API router composition
│   ├── api_routes/             # Split endpoint modules by domain
│   ├── models.py               # Pydantic schemas
│   ├── syllabus_parser.py      # Syllabus PDF extraction + events
│   ├── priority_engine.py      # Priority scoring + daily plan builder
│   ├── break_detector.py       # Schedule gap detection
│   ├── time_fitter.py          # Fits study items into break lengths
│   ├── quiz_generator.py       # Quiz creation logic
│   ├── utils.py                # Priority scoring + helpers
│   ├── requirements.txt
│   └── .env (ignored)
│
├── frontend/
│   ├── src/
│   │   ├── pages/
│   │   │   ├── Landing.tsx
│   │   │   ├── Onboarding.tsx
│   │   │   ├── onboarding/     # Split onboarding step components
│   │   │   └── Dashboard.tsx
│   │   ├── components/
│   │   ├── App.tsx
│   │   └── index.css
│   ├── package.json
│   └── tailwind.config.js
│
├── bot/
│   ├── telegram_bot.py         # Bot commands
│   ├── message_formatter.py    # Message templates
│   ├── requirements.txt
│   └── .env (ignored)
│
├── .env.example                # Copy to .env
├── .gitignore
└── README.md (this file)

---

Environment Variables

Copy .env.example to .env and fill in:

# Canvas
CANVAS_API_KEY=your_canvas_token_here
CANVAS_API_URL=https://your-school.instructure.com

# Supabase
SUPABASE_URL=https://xxxxx.supabase.co
SUPABASE_KEY=your_anon_key_here

# Telegram
TELEGRAM_BOT_TOKEN=123456:ABCdef-1234567890abcdefgh

# OpenAI (for quiz generation)
OPENAI_API_KEY=sk-...

# Frontend
REACT_APP_API_URL=http://localhost:8000

---

Database Schema (Supabase)

users

id, name, email, canvas_token, telegram_chat_id, created_at

courses

id, user_id, canvas_course_id, course_name, created_at

assignments

id, user_id, course_id, title, due_date, type, points, status, created_at

syllabus_events

id, user_id, course_id, title, event_type (exam/quiz/deadline), event_date, topic, created_at

schedules

id, user_id, day_of_week, start_time, end_time, class_name, created_at

study_recommendations

id, user_id, date, topic, reason, time_block, created_at

---

API Endpoints

Method	Endpoint	Purpose
POST	/connect/canvas	Validate & store Canvas token
GET	/courses	Get user's courses
GET	/today-plan	Get today's study plan
GET	/break-recommendation	Get next break suggestion
POST	/upload/syllabus	Upload & parse syllabus PDF
POST	/schedule	Save class schedule
POST	/telegram/send-morning-plan	Trigger morning message

---

Telegram Bot Commands

Command	Purpose
/start	Welcome & onboarding link
/today	Show today's plan
/nextquiz	When's the next quiz?
/weakest	What are my weak areas?
/quizme	Get a quick quiz

---

Example Output

Morning Message

Good morning! Here's your plan for today:

• 10:30–11:00 AM: Review STAT 200 hypothesis testing
• 1:20–1:35 PM: Quick CMPSC 131 practice
• 3:15–3:35 PM: Review Cyber notes for upcoming quiz
• 8:00–8:30 PM: Finish reading for AI-100

Top priority today: STAT 200 midterm is in 2 days.

Break Message

You have 25 minutes before your next class.
Best lesson to review: Hypothesis testing p-values
Reason: Upcoming exam in 2 days + low recent quiz score

Quick Quiz
1. What does a small p-value suggest?
   A) strong evidence against null
   B) null is proven true
   C) sample size is small

Reply with A, B, or C.

---

Tech Stack

Layer	Tech	Why
Backend	FastAPI	Fast, modern, great for AI apps
Database	Supabase + Postgres	Real-time, free tier
Frontend	React + Tailwind	Clean, fast, responsive
Bot	python-telegram-bot	Easy to use
PDF Parsing	PyPDF2	Simple, reliable
AI	OpenAI API	Quiz generation, summarization
Canvas	canvasapi	Official wrapper

---

Risks & Backups

Risk	Backup
Canvas API issues	Use mocked Canvas JSON data
Syllabus parsing messy	Support structured text input
Telegram timing hard	Manual trigger buttons
Weak-subject detection complex	Grade-by-course fallback

---

See Also

• Canvas API Docs: https://canvas.instructure.com/doc/api/
• Supabase Docs: https://supabase.com/docs
• Telegram Bot API: https://core.telegram.org/bots/api