Brain Squared when logged in
Brain Squared when logged out
Browsing Timeline Dashboard
Productive Hours
Qdrant Database
Brain Squared chat when prompted about react errors

🧠 Brain Squared

💡 Inspiration

We browse hundreds of sites every week — articles, docs, blogs — but when we try to recall something valuable we read, it’s nearly impossible to find it again. Browser history only works if you remember exact words, URLs, or page titles.
We wanted to fix that by creating a “second brain” for your browsing: a tool that remembers what you read and helps you find it again through meaning, not just keywords.

🚀 What It Does

Brain Squared automatically captures your browsing history from Chrome, stores it securely, and turns it into a searchable knowledge base powered by AI semantic search.

Key Features

🔁 Automatic Syncing – History syncs every 5 minutes automatically
🔍 Semantic Search – Find pages by meaning, not exact text
🧠 Smart Summaries – AI-generated TL;DRs for every page
🧩 Memory Graph View – Visualize your browsing knowledge as an interactive network
📱 Cross-Device Access – View and search your synced history from any device
🔐 Auth0 Login – Secure authentication across all devices
📊 Browsing Stats Dashboard – Retro cyberpunk-themed UI with live insights
🌐 Multi-Browser Support – Chrome, Edge, Firefox, Safari (in progress)
💬 AI Chat Search – Query your browsing history conversationally (“What was that article about local embeddings?”)

Example: Searching “React performance tips” finds “Optimizing React Apps” — even if you don’t remember the exact title.

⚙️ How We Built It

🖥️ Frontend

Next.js 14 + TypeScript + Tailwind CSS
Auth0 integration
Retro/cyberpunk aesthetic and responsive design
Recharts + Framer Motion for dynamic stats visualization

🧩 Backend

Node.js + Express + TypeScript
Qdrant Vector Database for semantic search
Xenova Transformers for local embedding generation
LangChain for summarization and semantic query handling
JWT authentication with Auth0

🔌 Chrome Extension

Manifest V3 service worker
Chrome Alarms API for scheduled syncing (every 5 minutes)
Chrome Storage API for token management
Manual and automatic history sync
Background script for summarization and embedding generation

☁️ Infrastructure

Qdrant Cloud for vector storage
Auth0 for authentication
Local development on localhost:3000 (web) and localhost:3001 (backend)

🧩 Challenges We Ran Into

Implementing JWT authentication between extension and backend
Fixing “Invalid Compact JWS” / “jwt malformed” errors from Auth0 audience mismatch
Handling large data batches efficiently (500+ history items per sync)
Building performant real-time summaries while keeping everything local
Designing the graph-based visualization without lag on large datasets
Balancing privacy with semantic embedding generation