A multi-strategy trading bot that identifies pricing inefficiencies in prediction markets. Combines BTC 5-minute microstructure analysis with ensemble weather forecasting to trade on Kalshi and Polymarket. Features a professional React dashboard.
100% free to run - No paid APIs, no subscriptions. All data sources are free. Kalshi API key optional for Kalshi markets.
Scans Polymarket BTC 5-minute Up/Down markets every 60 seconds. Uses real-time 1-minute candle data from Coinbase/Kraken/Binance to compute RSI, momentum, VWAP deviation, SMA crossover, and market skew as a weighted composite signal. Trades when edge > 2%.
Scans weather temperature markets on Kalshi (KXHIGH series) and Polymarket every 5 minutes. Uses 31-member GFS ensemble forecasts from Open-Meteo to estimate the probability of temperature thresholds being exceeded. Trades when edge > 8%. Kalshi markets are auto-discovered via the KXHIGHNY, KXHIGHCHI, KXHIGHMIA, KXHIGHLAX, KXHIGHDEN series tickers.
- BTC Microstructure Analysis - RSI, momentum (1m/5m/15m), VWAP, SMA crossover from real candle data
- Ensemble Weather Forecasting - 31-member GFS ensemble from Open-Meteo for probabilistic temperature predictions
- Multi-Platform Trading - Trades weather markets on both Kalshi (KXHIGH series) and Polymarket simultaneously
- Edge Detection - Identifies mispriced markets across both strategies and platforms
- Kelly Criterion Sizing - Fractional Kelly (15%) position sizing with per-trade caps
- Signal Calibration - Tracks predictions vs outcomes with Brier score
- Professional Dashboard - React 3-column dashboard with real-time updates
- Simulation Mode - Paper trading with virtual bankroll tracking and equity curves
cd kalshi-trading-bot
# Create virtual environment
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
# Install dependencies
pip install -r requirements.txt
# Run the backend
uvicorn backend.api.main:app --reload --port 8000Backend will be at: http://localhost:8000 API docs at: http://localhost:8000/docs
cd frontend
# Install dependencies
npm install
# Run the frontend
npm run devFrontend will be at: http://localhost:5173
┌──────────────────────────────────────────────────────────────────┐
│ FRONTEND │
│ React + TypeScript + TanStack Query + Tailwind │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ │
│ │Indicators│ │ Weather │ │ Signals │ │ Trades │ │
│ │ + Chart │ │ Panel │ │ Table │ │ Table │ │
│ └──────────┘ └──────────┘ └──────────┘ └──────────┘ │
└──────────────────────────────────────────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────────────┐
│ BACKEND │
│ FastAPI + Python + SQLite + APScheduler │
│ ┌───────────┐ ┌───────────┐ ┌───────────┐ ┌───────────┐ │
│ │ BTC │ │ Weather │ │ Signal │ │Settlement │ │
│ │ Signals │ │ Signals │ │ Scheduler │ │ Engine │ │
│ └───────────┘ └───────────┘ └───────────┘ └───────────┘ │
└──────────────────────────────────────────────────────────────────┘
│
▼
┌──────────────────────────────────────────────────────────────────┐
│ DATA SOURCES │
│ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌──────────┐ ┌────────┐ │
│ │Coinbase/ │ │Open-Meteo│ │ NWS │ │Polymarket│ │ Kalshi │ │
│ │Kraken/ │ │ Ensemble │ │ API │ │Gamma API │ │ API │ │
│ │Binance │ │ (GFS) │ │ │ │ │ │(KXHIGH)│ │
│ └──────────┘ └──────────┘ └──────────┘ └──────────┘ └────────┘ │
└──────────────────────────────────────────────────────────────────┘
- Fetch 60 one-minute candles from Coinbase/Kraken/Binance (fallback chain)
- Compute 5 indicators: RSI(14), Momentum(1m/5m/15m), VWAP deviation, SMA crossover, Market skew
- Convergence filter: require 2+ of 4 indicators to agree
- Weighted composite -> model UP probability (0.35-0.65 range)
- Compare to Polymarket prices, trade the side with higher edge
- Fetch open weather markets from Kalshi (KXHIGH series, RSA-PSS auth) and Polymarket (Gamma API)
- Fetch 31-member GFS ensemble forecasts from Open-Meteo
- Count fraction of members above/below the market's temperature threshold
- That fraction = model probability (e.g., 28/31 members above 70F = 90% probability)
- Compare to market price on either platform, trade when edge > 8%
- Confidence = ensemble agreement (how one-sided the 31 members are)
edge = model_probability - market_probability
BTC signals require |edge| > 2%. Weather signals require |edge| > 8%.
kelly = (win_prob * odds - lose_prob) / odds
position_size = kelly * 0.15 * bankroll
Capped at 5% of bankroll and $75 (BTC) or $100 (Weather) per trade.
| Endpoint | Method | Description |
|---|---|---|
/api/dashboard |
GET | All dashboard data in one call |
/api/btc/price |
GET | Current BTC price + momentum |
/api/btc/windows |
GET | Active BTC 5-min windows |
/api/signals |
GET | Current BTC trading signals |
/api/signals/actionable |
GET | BTC signals above threshold |
/api/kalshi/status |
GET | Kalshi API auth status + balance |
/api/weather/forecasts |
GET | Ensemble forecasts for all cities |
/api/weather/markets |
GET | Weather markets (Kalshi + Polymarket) |
/api/weather/signals |
GET | Weather trading signals (both platforms) |
/api/trades |
GET | Trade history |
/api/stats |
GET | Bot statistics |
/api/calibration |
GET | Signal calibration data |
/api/run-scan |
POST | Trigger BTC + weather scan |
/api/simulate-trade |
POST | Simulate a BTC trade |
/api/settle-trades |
POST | Check settlements |
/api/bot/start |
POST | Start trading |
/api/bot/stop |
POST | Pause trading |
/api/bot/reset |
POST | Reset all trades |
/api/events |
GET | Event log |
/ws/events |
WS | Real-time event stream |
All settings in backend/config.py, overridable via environment variables:
| Setting | Default | Description |
|---|---|---|
SCAN_INTERVAL_SECONDS |
60 | BTC scan frequency |
MIN_EDGE_THRESHOLD |
0.02 | Minimum edge (2%) |
MAX_ENTRY_PRICE |
0.55 | Max entry price (55c) |
MAX_TRADE_SIZE |
75.0 | Max $ per BTC trade |
KELLY_FRACTION |
0.15 | Fractional Kelly multiplier |
| Setting | Default | Description |
|---|---|---|
KALSHI_API_KEY_ID |
None | Kalshi API key ID |
KALSHI_PRIVATE_KEY_PATH |
None | Path to RSA private key PEM file |
KALSHI_ENABLED |
True | Enable/disable Kalshi market fetching |
| Setting | Default | Description |
|---|---|---|
WEATHER_ENABLED |
True | Enable/disable weather trading |
WEATHER_SCAN_INTERVAL_SECONDS |
300 | Weather scan frequency (5 min) |
WEATHER_MIN_EDGE_THRESHOLD |
0.08 | Minimum edge (8%) |
WEATHER_MAX_ENTRY_PRICE |
0.70 | Max entry price (70c) |
WEATHER_MAX_TRADE_SIZE |
100.0 | Max $ per weather trade |
WEATHER_CITIES |
nyc,chicago,miami,los_angeles,denver | Cities to track |
| Setting | Default | Description |
|---|---|---|
DAILY_LOSS_LIMIT |
300.0 | Daily loss circuit breaker |
MAX_TOTAL_PENDING_TRADES |
20 | Max open positions |
INITIAL_BANKROLL |
10000.0 | Starting paper bankroll |
| City | Station | Tracked |
|---|---|---|
| New York | KNYC | Default |
| Chicago | KORD | Default |
| Miami | KMIA | Default |
| Los Angeles | KLAX | Default |
| Denver | KDEN | Default |
Add more cities by editing WEATHER_CITIES in config and adding entries to CITY_CONFIG in backend/data/weather.py.
| Source | Data | Used For | Auth |
|---|---|---|---|
| Coinbase | BTC 1-min candles | BTC microstructure | None |
| Kraken | BTC 1-min candles | BTC fallback | None |
| Binance | BTC 1-min candles | BTC fallback | None |
| Open-Meteo | GFS Ensemble (31 members) | Weather probability | None |
| NWS API | Observed temperatures | Weather settlement | None |
| Polymarket | Market prices + resolution | Both strategies | None |
| Kalshi | Weather temperature markets (KXHIGH) | Weather strategy | RSA key |
kalshi-trading-bot/
├── backend/
│ ├── api/
│ │ └── main.py # FastAPI routes + dashboard
│ ├── core/
│ │ ├── signals.py # BTC signal generation
│ │ ├── weather_signals.py # Weather signal generation
│ │ ├── scheduler.py # Background jobs (BTC + weather)
│ │ └── settlement.py # Trade settlement (routes by market_type)
│ ├── data/
│ │ ├── btc_markets.py # Polymarket BTC market fetcher
│ │ ├── crypto.py # BTC price + microstructure
│ │ ├── kalshi_client.py # Kalshi API client (RSA-PSS auth)
│ │ ├── kalshi_markets.py # Kalshi weather market fetcher (KXHIGH)
│ │ ├── weather.py # Open-Meteo ensemble + NWS observations
│ │ ├── weather_markets.py # Polymarket weather market fetcher
│ │ └── markets.py # Generic market wrapper
│ ├── models/
│ │ └── database.py # SQLAlchemy models (market_type column)
│ └── config.py # All settings (BTC + weather)
├── frontend/
│ ├── src/
│ │ ├── components/
│ │ │ ├── GlobeView.tsx # 3D globe with city markers
│ │ │ ├── EdgeDistribution.tsx # Edge distribution chart
│ │ │ ├── MicrostructurePanel.tsx # RSI gauge + indicator meters
│ │ │ ├── WeatherPanel.tsx # Weather forecasts per city
│ │ │ ├── CalibrationPanel.tsx # Prediction accuracy tracking
│ │ │ ├── StatsCards.tsx # Performance metrics
│ │ │ ├── SignalsTable.tsx # BTC + Weather signals combined
│ │ │ ├── TradesTable.tsx # Trade history
│ │ │ ├── EquityChart.tsx # P&L chart
│ │ │ └── Terminal.tsx # Event log + controls
│ │ ├── App.tsx # 3-column grid dashboard
│ │ ├── api.ts # API client
│ │ └── types.ts # TypeScript interfaces
│ └── package.json
├── requirements.txt
├── run.py
└── README.md
This is a simulation tool for educational purposes. It does not place real trades or use real money. Past performance in simulation does not guarantee future results. Prediction markets involve risk of loss.
MIT - do whatever you want with it.