HackStack

Project Info

Quizzly - AI-Powered Educational Quiz Generator

Devpost

Inspiration

Picture this: You're watching an educational video about machine learning, taking notes furiously, but when the video ends, you realize you've absorbed almost nothing. Sound familiar? We've all been there - the endless scroll of educational content that promises to teach us everything, but leaves us with nothing but a sense of information overload and the nagging feeling that we're not actually learning. The problem hit us during our own study sessions. We'd watch hours of educational content, convinced we were learning, only to discover we couldn't recall the key concepts when it mattered most. Traditional video learning is passive, one-size-fits-all, and frankly, it's not working for most people. We wanted to change that. The breakthrough came when we realized that learning isn't just about consuming content - it's about active engagement, personalized feedback, and adaptive challenges that grow with you. What if we could transform any educational video into an intelligent, personalized learning experience that actually helps you retain and understand the material?

What it does

Quizzly is more than just another quiz app - it's your personal AI learning companion that transforms any educational video into an intelligent, adaptive learning experience. Here's the magic: upload any educational video, and Quizzly uses advanced AI to analyze the content, understand the key concepts, and generate personalized quizzes that adapt to your learning style and pace. But here's where it gets interesting. Quizzly doesn't just throw random questions at you. It watches how you learn, understands your strengths and weaknesses, and automatically adjusts the difficulty to keep you challenged but not overwhelmed. Get a question wrong? Instead of just showing you the answer, Quizzly generates a personalized explanation that helps you understand not just what you got wrong, but why you got it wrong and how to think about it differently. The system learns from your performance patterns, suggesting related topics to explore, identifying skill gaps, and even recommending the perfect next video to watch based on your learning journey. It's like having a personal tutor who knows exactly how you learn and what you need to succeed.

How we built it

We started with the Reka Vision API to analyze video content and extract meaningful educational concepts. This wasn't just about transcribing speech, we needed the AI to understand context, identify key learning objectives, and recognize the educational value of different segments. The real challenge came with the adaptive difficulty system. We spent hours analyzing learning patterns, studying how people progress through educational content, and building algorithms that could predict the optimal challenge level for each user. The system tracks your recent performance, identifies trends, and automatically adjusts question difficulty - if you're acing everything, it ramps up the challenge; if you're struggling, it provides more foundational questions to build your confidence. The personalized explanations feature was our secret sauce. When you get a question wrong, our AI doesn't just tell you the correct answer - it analyzes your specific mistake, considers the context of the video content, and generates a custom explanation that addresses your particular misunderstanding. It's like having a tutor who can see exactly where your thinking went wrong and guide you back to the right path. The smart recommendations engine was where we really pushed the boundaries. By analyzing your learning patterns, topic preferences, and performance trends, the system can suggest not just what to learn next, but how to learn it. It identifies your learning style, recommends related topics that build on your current knowledge, and even suggests when you might be ready to tackle more advanced concepts.

Challenges we ran into

The biggest challenge was making the AI truly intelligent, not just smart. Early versions of Quizzly would generate questions, but they felt robotic and disconnected from the actual learning experience. We spent hours refining the prompts and algorithms to make the AI understand not just what was said in the video, but what the learner needed to understand. The personalized explanations feature was technically complex - we needed the AI to not just identify wrong answers, but understand the reasoning behind those wrong answers and provide targeted feedback. This required sophisticated natural language processing and a deep understanding of common learning misconceptions. Performance was another challenge. With all the AI processing, real-time analytics, and adaptive features, we had to carefully optimize every component to ensure a smooth user experience. The system needed to feel instant and responsive while doing incredibly complex work behind the scenes.

Accomplishments we're proud of

We're incredibly proud of creating a system that truly understands how people learn. The adaptive difficulty feature works so well that users often don't even notice it's happening. They just feel like the content is perfectly matched to their level. The personalized explanations have received amazing feedback, with users saying they finally understand concepts they've been struggling with for years. The analytics dashboard is something we're particularly proud of. It doesn't just show numbers, it tells a story about the user's learning journey, highlighting progress, identifying patterns, and providing actionable insights that actually help people learn better. But what we're most proud of is the impact we're seeing. Users are reporting that they're retaining information better, learning faster, and actually enjoying the process of education again. We've created something that doesn't just test knowledge. It builds it.

What we learned

Building Quizzly taught us that the future of education isn't about replacing human teachers. It's about augmenting human learning with intelligent technology that adapts to individual needs. We learned that effective learning isn't just about content delivery, but about creating the right conditions for understanding to emerge. We discovered that personalization in education isn't just about different difficulty levels - it's about understanding how each person thinks, learns, and processes information. The AI needed to be empathetic, not just intelligent. The technical challenges taught us the importance of user experience in AI applications. The most sophisticated AI in the world is useless if it doesn't feel natural and helpful to the user. We learned to prioritize the human experience over technical complexity.

What's next

The learning revolution is just beginning. We're working on expanding Quizzly to support multiple languages, making it accessible to learners worldwide. We're developing mobile apps that bring the power of adaptive learning to your phone, so you can learn anywhere, anytime. We're exploring partnerships with educational institutions to integrate Quizzly into formal learning environments, creating a bridge between traditional education and AI-powered personalized learning. The goal is to make this technology available to every student, regardless of their background or resources. The future is collaborative. We're building features that allow learners to study together, share insights, and learn from each other's perspectives. Imagine a study group where the AI can facilitate discussions, suggest topics, and help everyone learn more effectively. We're making learning measurable. Advanced analytics will help learners understand not just what they know, but how they learn best. We're developing predictive models that can identify when someone is about to struggle with a concept and provide proactive support. The vision is clear: a world where every learner has access to personalized, adaptive, intelligent education that helps them reach their full potential. Quizzly is just the beginning of that journey.

Analysis

Compare with all teams

View

Metric

Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.

Technology

Found in codeClaimed only
  • CSSIn code
  • FlaskIn code
  • HTMLIn code
  • PythonIn code
  • JavaScriptClaimed

4 of 5 appear in the indexed code. 1 claimed on Devpost could not be matched to code, which may simply mean the tool leaves no trace in the repository.

AI coding agents

No AI coding agent signals were found in this repository.

Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.

Codebase size

Source size

263 KB

Source files

9

Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.

0 stars