Featured Project
QuestionCraft
AI-Powered Quiz Generation System
Completed
Status
2024
Year
Overview
This project aims to build an AI-powered quiz generation system that improves student learning by automatically generating and evaluating questions from educational content provided as a pdf. Using advanced LLM models like Nous-Research Heremes-3 and embedding model e5-small, the system creates contextually relevant question-answer pairs to support active recall and deeper understanding. It helps students prepare for exams by offering personalized question sets and supports interactive learning through a user-friendly platform built with React.
Key Features
OCR from PDFs: Extracts text from PDF documents using Optical Character Recognition to enable processing of scanned or image-based content.
Automatic Topic Identification: Automatically detects and categorizes key topics within the uploaded content to organize questions effectively.
Question Customization: Allows users to modify the type, difficulty, or focus of generated questions to suit specific learning needs.
Dynamic Question Generation: Uses AI to generate diverse and contextually relevant questions based on the content provided.
Summary: Provides concise summaries of educational material to support quick understanding and revision.
Explanation: Offers detailed explanations for each question and answer to reinforce learning and clarify concepts.
Source: Links each question back to its source content within the document for easy reference and context.
Fast Mode: Enables quicker generation by extracting text from the pdf instead of using OCR, suitable for text-based PDFs.
Intuitive User Interface: Designed with a clean and user-friendly layout to ensure ease of use for both students and educators.
Technology Stack
React.js
Tailwind CSS
FastAPI
Python
HuggingFace
LangChain
OpenRouter
PyPDFium2
PaddleOCR
Demo Video
Project Documents
Interested in this project?
Feel free to explore the live demo, check out the source code, or get in touch to discuss similar projects.