Kalshi Explore

Scans Kalshi prediction markets and compares prices to external data sources to identify potential mispricing opportunities, with multi-provider AI analysis for trade recommendations.

Status

Status	Active
Type	Personal Project

Links

• GitHub Repository

Overview

A comprehensive trading tool for Kalshi prediction markets that:

• Fetches real-time market data and compares to external sources (weather, sports, crypto)
• Identifies mispriced contracts with edge calculations
• Provides multi-provider AI analysis (Grok, OpenAI, Claude) with consensus voting
• Supports paper and live trading with risk management

Tech Stack

• Backend: Python (Flask API server)
• Frontend: React (Vite), TypeScript
• AI: Grok, OpenAI GPT, Claude
• Data Sources:
  • The Odds API (sports)
  • National Weather Service / Visual Crossing (weather)
  • CoinGecko (crypto)
  • RSS feeds (news)

---

Setup & Development

# Setup Python environment
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

# Configure API keys
cp .env.example .env
# Add: Kalshi, The Odds API, Visual Crossing, Grok, OpenAI, Claude keys

# Run CLI scanner
python main.py scan -v

# Run web dashboard
python -m web.app  # Flask on port 5050

# Or React frontend
cd frontend && npm install && npm run dev  # Port 3000

CLI Commands

Command	Description
python main.py scan	Single market scan
python main.py scan --watch	Continuous monitoring (60s refresh)
python main.py balance	View account balance
python main.py positions	View open positions
python main.py trade TICKER --side yes --qty 10 --price 25	Place limit order

Architecture

Data Fetchers

Source	Markets	Confidence
Sports (sportsbooks)	NFL, NBA, MLB, NHL	High
Weather (NWS/VC)	Temperature predictions	High
Crypto (CoinGecko)	BTC, ETH, SOL price targets	Medium-High
News (RSS)	Context for all markets	N/A

AI Analysis

Multi-provider consensus system:

• Grok (grok-4-1-fast-reasoning)
• OpenAI (gpt-5-mini)
• Claude (claude-sonnet-4)

Returns: trade/watch/skip recommendation with agreement level (unanimous/high/mixed)

Trading Safety

• Defaults to demo mode (KALSHI_ENV=demo)
• Confirmation prompt for production orders
• Risk limits: $50/order, $100/day loss limit, 100 contracts/market
• All trades logged to logs/trades.log

---

Handoff Notes

API Keys Required

• Kalshi API key + RSA private key for trading
• External data source keys (The Odds API, Visual Crossing)
• AI provider keys (Grok, OpenAI, Claude)