Inspiration
We wanted to explore biotech and looked into Chai Discovery, but couldn’t quite understand what they did, so we built something to fix that.
What It Does
Turns curiosity into visual knowledge maps powered by agentic AI search, making deep dives more organized, interactive, and fun.
How We Built It
Built with Next.js for the frontend and a Python backend handling agentic search, semantic clustering, and data processing.
Challenges We Faced
- Designing an intuitive UX for graph networks was tricky: too much complexity breaks usability.
- Optimizing agentic search performance was a technical challenge.
Accomplishments We’re Proud Of
- Achieved fast and responsive agentic search.
- Developed effective semantic clustering for organizing ideas.
What We Learned
How to design creative, user-friendly interfaces that make complex information approachable and useful.
What’s Next for Rabbithole
- Improving clustering algorithms for subtopics.
- Enhancing topic organization to better support deep, connected learning.
Built With
- ai
- mcp
- nextjs
- python

Log in or sign up for Devpost to join the conversation.