lavender

• 

  1-landing

• 

  2-loading

• 

  3-question-blank

• 

  4-question-with-hint

• 

  5-question-solved

• 

  6-finished

Inspiration

As students, we often struggle with organizing our study materials efficiently. The idea for this project was born from the need of a time saving automated tool that could transform raw study materials into structured questions, helping students study smarter.

What We Built

Our web application takes PDF notes and processes them using AI to generate a complete study set. The study set includes:

• AI-generated questions based on the content of the PDF.
• Hints to help guide the user without giving away the answer.
• Different types of answers, including multiple-choice, true/false, and short-answer formats.

How We Built It

Tech Stack

• Backend: Node.js, Express.js
• AI Integration: OpenAI API or any other LLM
• Data Handling: JSON format for structured study sets
• Frontend: Next.js

Architecture

1. The frontend sends a PDF file to the backend.
2. The backend extracts the text from the PDF.
3. AI processes the extracted text and generates study questions.
4. The API responds with a JSON object containing structured questions, hints, and answers.

Challenges We Faced

1. Handling Large PDFs

   • Some PDFs contained too much text, slowing down processing. We implemented strategies to summarize and extract key points efficiently.

2. Maintaining Formatting Consistency

   • Some PDFs had tables, images, or unusual formatting that made text extraction difficult. We had to implement preprocessing techniques to clean up the extracted text.

Future Improvements

• Interactive Flashcards: Add a feature where users can interact with the generated questions in a flashcard-style format.
• Multi-Language Support: Expand the tool to generate questions in multiple languages.
• More AI Fine-Tuning: Improve the accuracy and contextual relevance of the generated questions.

Built With

• express.js
• materialui
• multer
• next.js
• node.js
• openai
• pdfjs
• react.js