orbit

Inspiration

Our friends are UGC creators, and we've seen them struggle with the same challenge: creating content that consistently engages audiences and maintains high retention.

What it does

Orbit simulates how content spreads through social networks by simulating a society. Upload a video, and Orbit:

Analyzes your content using Google Gemini AI
Generates hundreds of realistic AI personas (complete with demographics, interests, and personalities)
Simulates their reactions through viewing, liking, commenting, sharing
Builds a dynamic social network showing how personas connect and influence each other
Models viral spread as content orbits through the network
Visualizes everything as an interactive 3D constellation where each persona becomes a star, sized and colored by engagement level

How we built it

Backend: Python, FastAPI, LangGraph for AI pipeline orchestration, Google Gemini AI for video analysis and persona simulation Frontend: Next.js, React, TypeScript, Three.js, Tailwind CSS

Challenges we ran into

Designing the LangGraph architecture to orchestrate video analysis, persona generation, network building, and viral simulation in the right sequence

What we learned

Visualization transforms data: The constellation metaphor isn't just pretty—it makes network analysis accessible and intuitive
Social dynamics are complex: Modeling how people influence each other required layered simulation (initial reactions + social influence reactions)