NeuroLearn AI

NeuroLearn AI is an adaptive studio that transforms how students master complex topics using multi-style explainers, D3 concept graphs, Feynman technique coaching, quizzes, and roadmaps.

Comment

Inspiration

Did you know that students forget over 70% of new information within a week if it isn't reinforced using active learning techniques? Traditional education often relies on passive reading and one-size-fits-all lectures.

We were inspired by the Ebbinghaus Forgetting Curve and cognitive science research which proves that people learn best when information is adapted to their personal learning style, visualized as connected concepts, evaluated through Feynman teaching, and tested adaptively. We built NeuroLearn AI to make deep mastery accessible, engaging, and personalized for every student on Earth.

What it does

NeuroLearn AI is a multi-modal adaptive learning studio that equips students with 5 specialized AI tools:

  1. Concept Explainer — Explains any topic in 5 tailored learning styles: ELI5 (analogies), Scholar (academic), Technical (deep mechanics), Storytelling (narratives), and Analogy-based.
  2. Interactive Knowledge Graph — Generates live, force-directed D3.js concept maps that visualize core ideas, prerequisites, and relationships with interactive node selection and zoom capabilities.
  3. Feynman Technique Coach — Evaluates a student's explanation in plain language against a strict scoring rubric (0–100%), highlighting specific knowledge gaps with feedback.
  4. Adaptive Quiz Studio — Generates dynamic MCQs tailored to the user's chosen topic and difficulty level (Easy, Medium, Hard) with auto-grading and a final Mastery Scorecard.
  5. Adaptive Study Roadmap — Creates custom curriculums tailored to any target timeframe with daily/weekly milestones and downloadable study plans.

How we built it

We designed NeuroLearn AI with a modern, zero-dependency, ultra-fast tech stack:

  • Frontend Interface: HTML5 and Vanilla CSS3 featuring glassmorphism, responsive CSS Grid layout, and dark cyber aesthetics.
  • Dynamic Logic: Vanilla JavaScript (ES6+, Strict Mode) built for high performance and smooth SPA navigation.
  • Data Visualizations: Integrated D3.js v7 for physics-based force-directed graph simulations and node drag interactions.
  • Markdown & Math: Marked.js for real-time rich text and markdown rendering.
  • AI Integration: Powered by the OpenRouter API (openrouter/auto live model routing) for fast, free-tier accessible multi-model LLM generation.

Challenges we ran into

  • Ensuring Structured JSON Output from LLMs: AI models sometimes wrap JSON outputs in extra text or markdown formatting. We implemented robust regex extraction and custom data normalizers (normalizeQuiz, normalizeGraph) to sanitize LLM responses before rendering.
  • Interactive D3.js Graph Layout: Balancing force simulation parameters (charge, link distance, collision detection) so nodes spread naturally without cluttering or overflowing the container bounds.
  • Zero Browser Scrollbar Layout: Locking the viewport grid (100vh) while allowing independent custom-scrollable panels for the sidebar and main studio area.

Accomplishments that we're proud of

  • Seamless 5-in-1 Studio: Created a unified, stateful interface where changing a topic seamlessly updates all 5 AI modules instantly.
  • Ultra-Fast & Serverless: Built the entire platform as a lightweight client-side SPA with zero heavy npm server dependencies needed to run locally.
  • Visual Polish: Crafted a premium Rose Gold & Dark Cyber design system with custom micro-animations and D3 canvas rendering.

What we learned

  • How to construct rigid system prompts for LLMs to enforce valid structured output formats.
  • Advanced D3.js force simulation, zoom event handlers, and dynamic SVG marker manipulation.
  • Cognitive learning frameworks (Feynman Technique, Active Recall, Spaced Repetition) and how to encode them into AI prompt engineering.

What's next for NeuroLearn AI

  • Flashcard & Anki Integration: Auto-generate downloadable Anki deck .apkg files directly from generated quizzes and concept maps.
  • Voice-Powered Feynman Coach: Allow students to explain concepts out loud using Speech-to-Text for a truly hands-free learning experience.
  • Collaborative Study Rooms: Real-time multi-user knowledge graph building for group study sessions. ## Inspiration

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for NeuroLearn AI

Built With

Share this project:

Updates

Submission history