Kalshi Trading Bot

Trading bot for Kalshi with a web dashboard to control it, manage triggers, and monitor activity. Uses SQLite for persistence (no separate database server).

Setup

1. Create a virtualenv and install dependencies

   python3 -m venv .venv
   source .venv/bin/activate
   pip install -r requirements.txt

2. Configure Kalshi API

   • Copy .env.example to .env
   • Get your API Key ID and private key from Kalshi account profile
   • Set KALSHI_API_KEY_ID and KALSHI_PRIVATE_KEY_PATH (or KALSHI_PRIVATE_KEY)
   • Use https://demo-api.kalshi.co/trade-api/v2 for testing, or https://api.elections.kalshi.com/trade-api/v2 for production

3. Run the dashboard

   python app.py

   Open http://127.0.0.1:5050 in your browser.

Project layout

File	Purpose
app.py	Flask entrypoint: serves dashboard and API, manages engine thread
routes.py	REST API: bot control, trigger CRUD, logs, portfolio, markets
bot_engine.py	Background loop: evaluates triggers, places orders via Kalshi
kalshi_client.py	Builds KalshiClient from .env credentials
db.py	SQLite schema and connection helpers
static/index.html	Dashboard UI (vanilla HTML/JS/CSS)
test_connection.py	Quick script to verify Kalshi API connectivity

API endpoints

Method	Path	Description
GET	/api/status	Bot running state
POST	/api/bot/start	Start the bot engine
POST	/api/bot/stop	Stop the bot engine
GET	/api/triggers	List all triggers
POST	/api/triggers	Create a trigger
PATCH	/api/triggers/<id>	Update a trigger
DELETE	/api/triggers/<id>	Delete a trigger
GET	/api/logs	Activity log (optional ?limit= and ?level=)
GET	/api/portfolio	Kalshi balance and positions
GET	/api/markets	Browse Kalshi markets

Trigger types

• price_above / price_below: fires when a market's price crosses a threshold (in cents). Config: market_ticker and/or series_ticker, threshold_cents, side (yes/no), optional contracts (default 1). Use series_ticker (e.g. KXETH-1H) to trade the current contract in that series (e.g. hourly ETH).

Triggers auto-disable after firing (one-shot). Re-enable from the dashboard.