Quillium

Quillium turns dense PDFs into quizzes, flashcards, and AI-powered study shorts using Google Gemini , so students learn actively, not passively.

Comment

Inspiration

Studying today often means dealing with long PDFs, dense notes, and static content that is hard to digest — especially during short study windows or last-minute revision.

For many students, this becomes even more challenging when:

  • The content is not in their native language
  • They don’t know how to self-test effectively
  • They lack personalized feedback while studying

We wanted to explore how modern AI tools like Google Gemini can turn passive reading into active learning — fast.

Quillium was inspired by a simple idea:

Learning should adapt to students, not force students to adapt to rigid study methods.

By combining document understanding with AI-powered quizzes, flashcards, and short-form explanations — across 50+ languages — Quillium aims to meet learners where they are.

What it does

Quillium is an AI-powered learning assistant that transforms static PDF documents into interactive study experiences.

With Quillium, users can:

  • Upload a PDF (notes, textbooks, slides)
  • Automatically generate MCQs using Gemini
  • Study with flashcards derived from those questions
  • Ask questions through an AI tutor (“Ask Quill”)
  • Track learning progress
  • Generate short-form study scripts and audio narration (“Study Shorts”)

Instead of rereading PDFs, learners actively quiz themselves, revise quickly, and learn in formats that match modern attention patterns.

How we built it

Backend

  • FastAPI (Python)
  • PyMuPDF for PDF text extraction
  • Google Gemini used for:
    • MCQ generation
    • Flashcard creation
    • Short-form explanation scripts
    • Ask Quill tutoring responses
  • OpenAI TTS for optional audio narration
  • Clean, modular REST APIs for each learning action

Frontend

  • Next.js (App Router) + React + TypeScript
  • Tailwind CSS with animated, cinematic UI
  • Dedicated sections for:
    • Upload
    • Quiz
    • Flashcards
    • Progress
    • Study Shorts
  • One-page navigation for fast interaction
  • LocalStorage-based progress tracking

AI Flow

PDF upload → text extraction → Gemini prompts → MCQs → flashcards → study shorts → progress tracking

Challenges we ran into

  • Extracting clean, structured text from different PDF formats
  • Ensuring AI-generated MCQs were meaningful, not repetitive
  • Keeping Gemini responses concise and student-friendly
  • Balancing advanced AI features with a beginner-friendly UI
  • Explaining complex content in short, engaging formats

We addressed these by refining prompts, adding validation checks, and testing with multiple document types and languages.

Accomplishments that we're proud of

  • Built a complete end-to-end AI learning prototype within the hackathon timeframe
  • Successfully integrated Google Gemini across multiple learning features
  • Converted PDFs into quizzes, flashcards, and short study scripts
  • Added multilingual learning support (50+ languages)
  • Designed a clean, engaging UI that encourages active learning

Most importantly, we built something that feels useful, not just impressive.

What we learned

  • AI works best when it supports learning, not replaces thinking
  • Prompt quality directly impacts educational value
  • Short-form learning improves engagement and retention
  • Clear UX matters as much as model capability
  • Rapid experimentation is essential in AI-first hackathons

What's next for Quillium

  • Adaptive quizzes based on learner performance
  • Captions and subtitles for Study Shorts
  • Audio-only learning mode for accessibility
  • Collaborative study rooms
  • Scalable deployment for students and educators

Quillium started as a hackathon prototype, but it has the potential to grow into a practical, inclusive learning companion powered by AI.

Team Name : Flare

   Members:  Jessy Kiruba  G,Faleesha Zaeen Zarshad

Built With

Share this project:

Updates