Orbital

Inspiration

Science today is fragmented - knowledge lives in isolated disciplines, slowing progress on the world’s hardest problems. Meanwhile, AI is reshaping fields like art and language, but science - the engine of human advancement - remains largely untouched by its full potential. I built Orbital to change that. Inspired by the dream of a truly unified scientific workspace, Orbital is designed to dissolve disciplinary boundaries, accelerate theory generation, and help humanity tackle open problems by giving every scientist - regardless of background - an AI-powered platform to explore, connect, and create new knowledge.

What it does

Orbital is a pan-scientific AGI-powered discovery platform that lets users explore, generate, and unify scientific knowledge across Physics, Biology, Chemistry, Mathematics, Computer Science, and beyond. It combines interactive theory maps, AI-driven theory generation, data analysis, and historical timelines into a single intuitive workspace. Users can visualize how theories connect, resolve contradictions, upload data for AI-assisted interpretation, and even collaborate with the AI to draft entirely new, testable scientific frameworks. Orbital helps researchers navigate the full landscape of human knowledge - and expand it.

How we built it

The Orbital app was built using a sophisticated stack of modern web technologies centered around React with TypeScript. The frontend leverages Vite as the build tool for fast development and optimized production builds. The UI is styled with Tailwind CSS for responsive design with a custom color scheme. Interactive theory maps are implemented using Vis.js network graphs with custom physics settings for engaging visualizations. Mathematical equations are rendered with KaTeX. The app uses React Router for navigation between different scientific disciplines (Physics, Biology, Chemistry, Mathematics, Computer Science, and Meta). For backend services, it integrates with Supabase for authentication and database functionality, while OpenAI's GPT-4 API powers the generative features that create novel scientific theories. ArXiv integration is handled through a Supabase Edge Function that serves as a proxy for fetching relevant scientific papers. The architecture is organized by discipline, with each having similar components (theory maps, data analysis, theory generation, reconciliation, and timelines). State management combines React Context for global state with localStorage for persisting generated theories, creating a comprehensive platform for exploring scientific knowledge across disciplines.

Challenges we ran into

Building Orbital presented several significant challenges. The most demanding was creating a cohesive cross-disciplinary knowledge framework that accurately represents relationships between theories across physics, biology, chemistry, mathematics, and computer science. Designing an intuitive visualization system for complex theory networks required extensive work with the Vis.js library to achieve the right physics settings and interactive elements.

Integrating real-time arXiv paper fetching through Supabase Edge Functions presented technical hurdles, particularly with XML parsing and rate limiting. The generative AI components required careful prompt engineering to produce scientifically plausible theories with appropriate mathematical notation. Ensuring consistent UI/UX across multiple scientific disciplines while maintaining discipline-specific terminology and visualization schemes was also challenging.

Accomplishments that we're proud of

I'm particularly proud of creating a truly interdisciplinary platform that bridges traditional academic silos. The meta-theoretical map visualizing connections across disciplines represents a novel approach to knowledge representation. The interactive theory network visualizations with custom physics settings create an engaging and intuitive way to explore complex scientific relationships.

The AI-powered theory generation system produces remarkably plausible scientific hypotheses complete with equations, predictions, and experimental proposals. We successfully implemented a real-time arXiv integration that fetches relevant papers for any theory in the system. The chat-based exploration of generated theories allows for deep dives into speculative science with contextual awareness. Finally, the consistent design language across disciplines creates a unified experience while respecting the unique characteristics of each field.

What we learned

Building Orbital showed me that AI isn’t just capable of summarizing knowledge — it can actively propose new scientific ideas, detect hidden connections between fields, and assist in resolving theoretical conflicts. We learned how powerful it is to visualize science as a living, interconnected map, and how essential cross-disciplinary tools are for modern discovery. Perhaps most importantly, I saw that the hunger for this exists — scientists and curious minds alike want AI to not replace them, but to amplify their creativity and accelerate their ability to generate breakthroughs.

What's next for Orbital

I believe Orbital can fundamentally change how science is done, and that vision doesn’t stop at a prototype. The next step is to transform Orbital into a fully-fledged startup, focused on making this AI-native scientific workspace available to every researcher, student, and knowledge-seeker on the planet. My goal is to democratize access to cutting-edge scientific tools, allowing anyone — from postdocs to independent thinkers — to explore, generate, and challenge scientific theories with AI as their co-pilot. Orbital won't just be an app — it will be a platform that empowers every human to become an active agent in the creation of new knowledge.

Built With

• and-vis.js-to-generate
• api
• arxiv
• css
• d3.js
• katex
• lucide
• openai
• papaparse
• react
• react-and-typescript-based-pan-scientific-agi-platform-that-uses-gpt-4
• router
• supabase
• tailwind
• typescript
• vis-network
• visualize
• vite