Polymarket Intel

Copy-intelligence system for Polymarket — Find smart wallets, detect consensus signals, paper trade, and autoresearch strategies.

What This Does

Instead of trying to predict markets (hard), we:

1. Analyze wallets — Score every Polymarket wallet by win rate, P&L, and volume
2. Detect consensus — When multiple smart wallets agree on a trade, that's a signal
3. Paper trade — Simulate trades on signals to test before risking real money
4. Autoresearch — Karpathy-style loop: try strategies, evaluate, keep/discard/modify

Architecture

polymarket-intel/
├── config.py               # All configuration
├── db.py                   # SQLite database layer
├── main.py                 # CLI entry point
├── api/
│   ├── gamma.py            # Gamma API (market data)
│   ├── trades.py           # Trades API (live trades)
│   └── scraper.py          # Combined scraper + storage
├── analysis/
│   ├── wallet_scorer.py    # Wallet scoring & leaderboard
│   └── correlation.py      # Multi-wallet consensus detection
├── simulation/
│   └── paper_trader.py     # Paper trading simulator
├── research/
│   └── autoresearch.py     # Strategy mutation + evaluation loop
├── alerts/
│   └── notifier.py         # Discord/Telegram/n8n alerts
└── data/                   # SQLite DB + collected data

Quick Start

cd ~/polymarket-intel
pip install -r requirements.txt

# Initialize database
python main.py init

# Run full pipeline (collect → score → signals → trade → research)
python main.py run

# Or step by step:
python main.py collect        # Fetch trade data
python main.py markets        # Sync market data
python main.py score          # Score all wallets
python main.py leaderboard    # Show top wallets
python main.py signals        # Detect consensus signals
python main.py paper trade    # Open paper trades on signals
python main.py paper status   # Show paper trading performance
python main.py paper resolve  # Resolve positions against market outcomes
python main.py research cycle # Run one autoresearch cycle
python main.py monitor        # Start continuous monitoring

How It Works

1. Wallet Intelligence

• Fetches trades from Polymarket's public API
• Tracks every wallet's trading history
• Calculates win rate, P&L, and composite score
• Classifies wallets: directional bettor, market maker, etc.

2. Signal Detection

• When 3+ high-scoring wallets BUY the same outcome = signal
• Weighted by wallet score, consensus strength, volume
• Signals stored in DB for paper trading and analysis

3. Paper Trading

• Virtual trades on detected signals
• Tracks balance, P&L, win rate
• Auto-resolve when markets close
• No real money at risk

4. Autoresearch

• Starts with 5 base strategies (conservative → aggressive)
• Evaluates each by Sharpe ratio, win rate, PnL
• Mutates best strategies, discards worst
• Over time, finds optimal strategy parameters

Configuration

Edit config.py or set environment variables:

export DISCORD_WEBHOOK_URL="https://discord.com/api/webhooks/..."
export TELEGRAM_BOT_TOKEN="your-bot-token"
export TELEGRAM_CHAT_ID="your-chat-id"
export N8N_WEBHOOK_URL="https://your-n8n.com/webhook/..."
export FIRECRAWL_API_KEY="your-key"

API Endpoints Used

Endpoint	Auth	Purpose
gamma-api.polymarket.com	None	Market data, resolution
strapi-matic.polymarket.com/trades	None	Live trade feed

No API keys needed for core functionality. Trade feed is intermittent — scraper has retry logic.

Data Flow

Polymarket APIs
      ↓
  [Scraper] → SQLite DB
      ↓
  [Wallet Scorer] → Wallet scores
      ↓
  [Correlation Engine] → Signals
      ↓
  [Paper Trader] → Simulated trades
      ↓
  [AutoResearch] → Strategy optimization
      ↓
  [Notifier] → Discord/Telegram/n8n alerts