Binance trading bot: API keys, code and a safe setup

Step 1 — create restricted API keys

In Binance, go to API Management and create a key. The security settings here matter more than any line of code:

Step 2 — connect with ccxt

python · binance_connect.pyimport ccxt, os
ex = ccxt.binance({
    'apiKey': os.environ['BINANCE_KEY'],
    'secret': os.environ['BINANCE_SECRET'],
    'enableRateLimit': True,        # ccxt throttles for you
})
ex.set_sandbox_mode(True)              # testnet
bal = ex.fetch_balance()
print(bal['USDT']['free'])

Step 3 — place a spot order

python · binance_order.py# market buy 0.001 BTC with the signal from your strategy
if signal() == 'buy':
    order = ex.create_market_buy_order('BTC/USDT', 0.001)
    print(order['id'], order['status'])

Where does signal() come from? Test the exact rule on our backtester with BTC and ETH series first — if it doesn't beat buy-and-hold after fees there, it won't on Binance either.

Step 4 — respect rate limits

Binance uses a weight system: every endpoint costs “weight” against a per-minute budget, and breaching it triggers a temporary IP ban. Set enableRateLimit: True in ccxt, batch requests, and poll the slowest timeframe your strategy allows. A 1-hour bot needs to call the API far less than a 1-minute scalper.

Step 5 — testnet before mainnet

Binance offers a full Spot Testnet with fake balances. Run your bot there for weeks: confirm orders fill, errors are handled, and the equity curve roughly matches your backtest before risking a cent.

Go-live checklist

1. Strategy beats buy-and-hold after fees on the backtester.
2. Keys are trade-only, no withdrawals, IP-whitelisted.
3. Tested for weeks on the Spot Testnet.
4. Position size from the calculator, risk ≤1%.
5. Max-drawdown kill switch wired in.

Frequently asked questions