PolyLens is an AI-powered intelligence platform designed to decode prediction markets. It bridges the gap between raw Polymarket odds and real-world context using high-speed research agents and LLM-driven synthesis.

<video src"https://www.youtube.com/watch?v=dQw4w9WgXcQ&list=RDdQw4w9WgXcQ&start_radio=1" controls width="100%">

PolyLens is built for speed and precision, using a modern hybrid architecture:

• Framework: Next.js 15+ (App Router)
• Language: TypeScript
• Styling: Tailwind CSS 4.0
• Runtime: Node.js / Edge Runtime

• LLM: Google Gemini 1.5 Flash (Ultra-fast reasoning)
• Deep Research: Custom synthesis pipeline that merges pre-cached news citations with live market data.
• Agentic Layer: Standalone Python research agent using browser-use for deep-web scraping.

• Markets: Polymarket Gamma API
• Real-time News: Google News RSS & direct citations.
• Community Pulse: Polymarket event comments and sentiment analysis.

PolySearch/
├── polylens/               # Main Next.js Web Application
│   ├── src/app/            # App Router & API Routes
│   ├── src/components/     # UI Components (Results, Leaderboards, Gauges)
│   └── backend/            # Research Agent Engine
│       └── polymarket_agent.py  # Heavy-duty Python Research Agent
├── .env.example            # Template for environment variables
└── README.md               # This project documentation

• Node.js: v18.18+ or v20+
• Python: v3.9+ (optional, for standalone agent)
• API Key: Google Gemini API Key

1. Clone the repo.
2. Create polylens/.env (use .env.example as a template).
3. Install frontend dependencies:

   cd polylens
   npm install

4. (Optional) Install Python agent dependencies:

   cd polylens/backend
   pip install langchain-google-genai browser-use playwright
   playwright install chromium

cd polylens
npm run dev

Navigate to http://localhost:3000.

When you search for a topic, PolyLens immediately fetches Polymarket odds and the latest relevant news articles via RSS. No AI delay.

Clicking "Generate Analysis" triggers a deep synthesis. Gemini parses the identified articles, cross-references them with market volume and liquidity, and builds a comprehensive Intelligence Dossier.

For heavy-duty research without the UI, you can run the Python agent directly:

python3 polylens/backend/polymarket_agent.py "Who will win the 2026 NBA Finals?" --json

Built with love for prediction market enthusiasts.