Docs · Getting Started

Run your first trading bot in 5 minutes

This guide walks you through installing WatchDog Bot, connecting your first exchange, and running a working Python trading bot — end to end, in under five minutes.

Download & install

Grab the desktop app from the homepage. Available for Windows and macOS. The installer is signed and around 200 MB — bundled Python runtime included, so there's nothing else to install.

→ Download WatchDog Bot

Sign in or create an account

First launch will open the auth screen. Use email + password, or sign in with Google. New accounts get a free trial automatically — no credit card needed.

Add your first exchange connection

Head to Settings → API Connections. Click Add Connection, pick your exchange, and paste in your API credentials.

Currently supported:

Kalshi (US prediction markets) — access key + private key file
Binance, Coinbase, Kraken (crypto) — API key + secret
Custom — any HTTP API with key + secret + base URL

Credentials are encrypted locally and synced to your account. They're never exposed to bot code directly — bots access them through the wd SDK.

Create your first bot

Click New Bot on the dashboard. You'll get a code editor. Either paste in a template, import from the library, or start from scratch. Here's a minimal example that just logs market data:

import time
import wd

conn = wd.connection("Kalshi")
wd.log.info("Bot started — connected to %s", conn.base_url)

for i in range(10):
    wd.log.info("Tick %d at %s", i, time.strftime("%H:%M:%S"))
    time.sleep(5)

wd.log.info("Bot finished")

No installs needed — import wd works out of the box. If your real bot uses requests, pandas, ccxt, or any other PyPI package, WatchDog Bot auto-installs it the first time it sees a ModuleNotFoundError. You don't need to maintain a requirements.txt.

Run it

Click Start. The status indicator turns green and live logs stream in real time below the editor. Each line shows level (INFO / WARN / ERROR), timestamp, and your message.

If something errors, you'll see a Fix with AI button. Click it — Claude reads your code plus the traceback and proposes a fix you can accept with one click.

That's it. You've got a running bot. From here, the rest is strategy. Check out the full Kalshi tutorial for a complete, runnable example with real order placement.

What's next

Build a complete Kalshi trading bot in Python
Python SDK reference — wd.log, wd.connection, wd.is_demo (coming soon)
Bot examples library — Kalshi market maker, simple momentum, news-driven bot (coming soon)
AI Fix guide — how it works, what it catches, when to trust it (coming soon)

Common first-time issues

"My bot isn't producing any logs"

Make sure you're using wd.log.info(...) not print(...). Both work, but wd.log produces structured output that the dashboard filters on. print output still appears, just unstructured.

"ModuleNotFoundError on launch"

The platform auto-installs missing dependencies and retries up to 3 times before giving up. If it still fails, the package name in your import may not match the PyPI name. For example, import cv2 → install opencv-python. Edit your code to use the right import name, save, and click Start again.

"My connection isn't showing up in wd.connection(...)"

The argument is the connection's slug. By default it matches the name you gave it (case-insensitive). If you named your connection "My Kalshi", call wd.connection("My Kalshi"). Spaces and case don't matter; punctuation does.

Ready to ship your first bot?

Free trial, no credit card. Download takes 2 minutes.

Start Free Trial →