# Project export: Introspect AI

This document was generated by HackStack to give an AI agent context about a hackathon project. Sections are labeled with their provenance; content marked as truncated was cut to keep this document small.

## Project metadata

- Hackathon: UC Berkeley AI Hackathon 2025
- Tagline: Visualize your ChatGPT and Claude conversations in 3D space
- Devpost: https://devpost.com/software/introspect-ai
- GitHub: not linked
- Video: https://www.youtube.com/embed/dMug-9h_eW4?enablejsapi=1&hl=en_US&rel=0&start=&version=3&wmode=transparent
- Team: contributor stats unavailable

## Devpost submission (written by the team)

### Overview

Your AI conversations, beautifully visualized in 3D space What It Does Introspect transforms your ChatGPT and Claude conversation history into an interactive 3D experience. Like Spotify Wrapped, but for your AI chats. Every message becomes a glowing point in space, clustered by meaning using UMAP and AI-powered topic detection. The Magic 3D Semantic Clustering Conversations with similar themes naturally group together using UMAP dimensionality reduction. AI-Powered Labels Claude auto-generates topic names like "Python & Data Analysis" or "Philosophy & Ethics". Multi-User Support Color-coded by user, with separate clusters and statistics per person. Interactive Exploration Smooth camera controls, zoom, search, filtering, and hover-to-view details. Real-Time Processing Live API backend, continuously refreshed as new data comes in. Tech Stack Frontend: React + Three.js + WebGL shaders for high-performance 3D rendering Backend: Flask + UMAP + Claude API + Voyage AI for embeddings + Groq for summarization and Chat Database: Supabase + pgvector for semantic search and storage ML Pipeline: SpaCy NLP → Voyage embeddings → UMAP 3D → KMeans Clustering + Claude SpaCy NLP → Voyage embeddings → UMAP 3D → KMeans Clustering + Claude Key Features Smart Parsing: Handles both Claude and ChatGPT exports, no manual cleanup needed Semantic Analysis: 512-dimensional embeddings mapped to 3D space Dynamic Clustering: Uses Claude Sonnet 4 to determine optimal cluster counts Performance Optimized: Embedding batches + GPU shader rendering Resilient: Generates mock data when APIs are down, so you can always demo The Experience Upload your conversations → Watch them bloom in 3D → See patterns you never noticed → Reflect, share, and explore your AI journey Built for AI power users, researchers, and curious minds.

## README (from the GitHub repository)

No README available.

## Detected evidence (automated analysis)

No repository was indexed for this project. Claimed technologies below could not be checked against code.
- Anthropic (technology) — claimed on Devpost, not found in the code
- TensorFlow (technology) — claimed on Devpost, not found in the code

## Codebase structure

No repository index available.

## Key source files

No repository index available; no source files included.