Inspiration
AI tools are making learning faster, but they are also making students increasingly dependent on generated answers. Many learners today can recognize information, summarize notes, or copy explanations from AI — but struggle to truly understand concepts deeply enough to explain them in simple words.
We were inspired by the Feynman Technique, which says:
“If you can’t explain something simply, you don’t understand it well enough.”
ClaryLight was created to solve a growing problem in modern education:
- declining critical thinking,
- passive AI-assisted learning,
- overreliance on instant answers,
- and the illusion of understanding.
Instead of building another AI tutor that simply gives answers, we wanted to build an AI system that challenges users to think clearly, identify gaps in their own understanding, and rebuild concepts from first principles.
Our goal was to transform AI from an answer machine into a clarity machine.
What it does
ClaryLight is an AI-powered learning platform that helps users validate and strengthen their understanding of any topic using the Feynman learning method.
Users explain concepts in their own words through typing or voice input, and the system analyzes their explanation to detect:
- missing concepts,
- misconceptions,
- logical gaps,
- and excessive academic jargon.
The platform then:
- generates personalized gap analysis,
- asks targeted follow-up questions,
- creates flashcards and learning resources,
- and visualizes the user’s understanding through a Growing Knowledge Map.
One of the core features is the Personal Tutor, an AI assistant that behaves like a curious student instead of a teacher. Rather than directly giving answers, it asks clarifying questions whenever the user’s explanation becomes unclear.
Another unique feature is the Jargon Jar, which gamifies learning by detecting complex terminology and encouraging users to explain difficult terms in simpler language.
ClaryLight also includes ELIF Mode (Explain Like I’m Five), which forces ultra-simple explanations and promotes genuine conceptual clarity over memorization.
How we built it
We built ClaryLight as a frontend-focused AI application using modern web technologies and lightweight architecture for rapid prototyping and scalability.
Tech Stack
- React + Vite for frontend development
- Tailwind CSS + Shadcn UI for responsive and clean UI design
- React Flow for the Growing Knowledge Map visualization
Gemini/OpenAI APIs for:
- gap analysis,
- tutor conversations,
- jargon detection,
- resource generation,
- and quiz generation
Web Speech API for real-time voice-to-text input
LocalStorage for session persistence and history tracking
Core System Design
The application workflow follows this structure:
- User selects a topic
- User performs a “brain dump” explanation
- AI analyzes the explanation
- Knowledge gaps and misconceptions are identified
- Personalized questions and resources are generated
- Understanding is visualized through the Growing Map
- User interacts with the Personal Tutor to refine concepts further
We designed the AI prompts carefully to ensure the tutor behaves like a confused learner instead of a traditional answer-giving chatbot.
Challenges we ran into
One of our biggest challenges was designing AI interactions that promote critical thinking instead of reducing it.
Most AI systems are optimized to provide direct answers instantly, but our challenge was:
- making AI ask better questions,
- detecting fuzzy understanding,
- and encouraging simplicity without frustrating the user.
Other technical challenges included:
- implementing reliable real-time voice-to-text functionality,
- designing the Growing Map visualization dynamically,
- balancing AI response quality with API costs and latency,
- and building a meaningful jargon detection system instead of simple keyword matching.
We also faced UX challenges while ensuring the platform remained educational, engaging, and not overwhelming for first-time users.
Another major challenge was maintaining feasibility within a hackathon timeline while still delivering a polished and impactful experience.
Accomplishments that we're proud of
We are proud that ClaryLight goes beyond being “just another AI study assistant.”
Some accomplishments we are especially proud of include:
- Creating an AI tutor that intentionally avoids giving direct answers
- Designing the Jargon Jar, a unique gamified clarity mechanism
- Building a system focused on understanding validation rather than memorization
- Successfully combining educational psychology with AI interaction design
- Making complex learning feel visual and interactive through the Growing Map
- Encouraging users to simplify ideas instead of hiding behind technical jargon
We are also proud that our solution addresses a real modern problem:
AI dependence reducing deep thinking and conceptual clarity.
What we learned
Through building ClaryLight, we learned that:
- educational AI should guide thinking, not replace it,
- simplicity is one of the strongest indicators of true understanding,
- and user experience is just as important as technical capability in learning platforms.
We also learned how prompt engineering can fundamentally change AI behavior. By changing the AI persona from “teacher” to “curious learner,” the quality of interactions became far more reflective and engaging.
On the technical side, we improved our understanding of:
- AI workflow orchestration,
- frontend state management,
- real-time voice processing,
- visualization systems,
- and educational product design.
Most importantly, we learned that AI should not only make learning easier — it should make learning deeper.
What's next for ClaryLight
Our future vision for ClaryLight is to evolve it into a complete cognitive learning ecosystem focused on deep understanding and independent thinking.
Planned future improvements include:
- adaptive mastery tracking,
- spaced repetition integration,
- collaborative study rooms,
- advanced concept dependency graphs,
- multilingual learning support,
- personalized learning pathways,
- and cloud synchronization across devices.
We also want to explore:
- AI-generated analogies,
- confidence heatmaps,
- deeper misconception analysis,
- and integration with academic curricula.
Long term, we envision ClaryLight becoming a platform that helps learners build clarity, curiosity, and critical thinking in an age increasingly dominated by passive AI consumption.
Our mission is simple:
“Use AI not to think for people — but to help people think better.”
Log in or sign up for Devpost to join the conversation.