An agentic system that tracks user knowledge to guide personalized learning journeys.
The Problem: As we offload more to AI, we're losing agency and understanding. AI boosts productivity but often reduces learning speed.
Our Solution: Long-context AI that knows what you know and guides you to optimal next learning steps.
The Result: Personalized learning paths that enhance rather than replace human understanding.
Winner of Lux Capital AI Agents & Infra Hackathon (August 2025) - Best Agent Hack, 2nd place overall
Created by Bright Liu and Fucheng Warren Zhu
# Install dependencies
npm install
# Create a learning workspace
python3 setup_workspace.py -d my-topic -t "machine learning"# Terminal 1: Start the monitoring server
npm start
# Terminal 2: Begin your learning session
cd workspaces/my-topic
claude -p "Let's start learning!"
# View progress at http://localhost:3001GraphMyMind creates a personalized learning environment that:
- Tracks Your Knowledge: Maintains a knowledge graph of what you understand
- Identifies Gaps: Compares your knowledge with comprehensive domain knowledge
- Guides Learning: Suggests optimal next concepts based on your current understanding
- Monitors Progress: Real-time visualization of your learning journey
- 🧠 Knowledge Graphs: Visual representation of concepts and relationships
- 📊 Gap Analysis: Identifies what you need to learn next
- 🎯 Personalized Paths: Adapts to your learning style and pace
- 📈 Progress Tracking: Real-time monitoring of your learning journey
- 🤖 Claude Integration: Seamless integration with Claude Code
GraphMyMind/
├── setup_workspace.py # Creates personalized learning workspaces
├── server.js # Real-time monitoring server
├── public/index.html # Learning progress UI
├── templates/ # Workspace templates
└── workspaces/ # Your learning sessions
- Python 3.8+
- Node.js 14+
- Claude Code CLI
MIT

