Inspiration
We were inspired by the pain of creating quizzes from scratch. As students and educators, we wanted to build a tool that could instantly transform any PDF study material into a ready-to-use quiz. We aimed to automate the learning process and save valuable time.
What it does
Quizlo.ai is a full-stack web application that generates custom quizzes from uploaded PDF documents. Users can sign in securely with their Google account via Firebase SSO. Once logged in, a user can upload a PDF, which is then processed by a custom GPT model to generate a multiple-choice quiz. The application provides a seamless and automated way to create personalized study tools.
How we built it
How we built it
We built the frontend using Next.js, React, and Tailwind CSS for a clean, responsive user interface. The backend is a Node.js and Express.js server, with MongoDB Atlas as the database to store user and quiz data. We integrated Firebase Authentication for secure user management and leveraged a fine-tuned GPT model to intelligently parse and generate quiz questions from the document content.
Challenges we ran into
We faced significant challenges with a number of issues. The most notable was a persistent CORS policy error that blocked communication between our frontend and backend. This required extensive debugging of our middleware and environment variable configurations. We also struggled with a tricky Firebase configuration issue and a frustrating merge conflict due to two team members editing the same database schema file.
Accomplishments
We're proud of successfully completing the entire end-to-end user flow. We managed to overcome the complex CORS hurdles, which were a major technical roadblock. We also successfully integrated a gemini model to intelligently generate the quizzes. Finally, despite the collaboration challenges, we successfully learned to debug and resolve complex Git issues to get our product working.
What's next for Quizlo.ai
We plan to expand Quizlo.ai with new features to enhance the user experience. We will add more question types, such as true/false and short answer. We also want to implement a user dashboard to track learning progress and a personalized recommendation system. Finally, we aim to extend the platform to support other document formats like Word and plain text files.
Log in or sign up for Devpost to join the conversation.