Home page
Add an AI Conversation and select which AI you chatted with
Human vs. AI parts are identified and put into clean UI for your partners to see your research
Add a parter as viewer or Space (Space is where you add AI conversations) editor
Semantically search through AI conversations you have put into your Space
Keyword search
View the topics you have talked about in 3D, which appear as fancy floating balls
Press on a topic to view relevant chat details and a relevant AI generated image to assist comprehension
Press on a topic to view relevant chat details and a relevant AI generated image to assist comprehension

Inspiration

Collaborative research today is messy and fragmented. Students and teams constantly switch between multiple AI tools, copy-paste responses, and lose track of sources across tabs and documents.

The classic product we are reimagining is the traditional research workspace, such as shared documents (e.g., Google Docs), notebooks, and browser tabs. These tools were not designed for AI-driven workflows, where information comes from multiple intelligent sources.

We wanted to rethink this entirely by creating a system that is structured, visual, and built specifically for multi-AI collaboration.

What it does

GenSpace is an AI-powered collaborative research workspace that transforms the traditional document into a structured, intelligent system.

Instead of scattered notes and tabs, users can paste conversations from different AI tools into a shared “Space,” where GenSpace:

Organizes information into a unified workspace
Automatically extracts and links sources
Visually distinguishes outputs from different AI models
Enables semantic (meaning-based) search across all content
Allows users to explore research in an immersive 3D environment with floating text and AI-generated visuals GenSpace turns fragmented AI outputs into a cohesive, explorable knowledge system.

How it works

Users create a Space and paste conversations from AI tools like ChatGPT, Claude, or Gemini. GenSpace then:

Parses the conversation into structured messages
Extracts any links and attaches them as sources
Tags each message with its AI origin
Stores everything in a shared database
Enables searching and exploration through both text and 3D views This allows teams to collaboratively build and navigate research in a much more intuitive way.

How I built it

GenSpace is built as a full-stack web application:
Frontend: Next.js + Tailwind CSS for a clean, responsive interface
Backend: Next.js API routes
Database: MongoDB for storing users, Spaces, and messages
Parsing System: Custom logic to structure pasted AI conversations
Source Extraction: Regex-based link detection and display
Search: Semantic search powered by embeddings
3D Exploration: Three.js / React Three Fiber for interactive research spaces
AI Integration: Gemini API for generating contextual images

Challenges I ran into

Parsing messy, inconsistent AI chat formats into structured data Ensuring summaries were meaningful rather than vague Designing a UI that balances simplicity with powerful features Managing performance with semantic search and 3D rendering Making multiple AI sources feel unified Keeping scope manageable within hackathon time constraints