Binance Auto-Trading Bot

Automated trading on Binance Spot and USDT-M Futures. Strategies: RSI, EMA crossover, combined (RSI + EMA), and Grid (spot only).

Defaults to testnet. Flip USE_TESTNET=false only after you've verified behavior on testnet.

Project layout

Binance/
├── src/
│   ├── main.py          # entrypoint
│   ├── exchange.py      # Binance client wrapper (spot + futures)
│   ├── strategy.py      # RSI / EMA / combined / grid
│   ├── risk.py          # SL/TP, sizing, daily loss kill-switch
│   ├── spot_bot.py      # spot engine
│   ├── futures_bot.py   # futures engine
│   └── logger.py
├── config.yaml          # strategy + risk parameters
├── .env.example         # API keys template
├── requirements.txt
├── Dockerfile
└── docker-compose.yml

1. Get testnet API keys

• Spot: https://testnet.binance.vision → "Generate HMAC API Key"
• Futures: https://testnet.binancefuture.com → log in with GitHub, then API Key

The two testnets are separate accounts with separate keys.

2. Setup

cp .env.example .env
# Edit .env and paste your testnet keys
python -m venv .venv
# Windows: .venv\Scripts\activate
# Unix:    source .venv/bin/activate
pip install -r requirements.txt

3. Edit config.yaml

Key settings:

Field	Meaning
risk.stop_loss_pct	SL as fraction of entry (0.05 = 5%)
risk.take_profit_pct	TP as fraction of entry
risk.order_size_usdt	notional per trade
risk.daily_loss_limit_pct	kill-switch threshold
spot.strategy	rsi / ema / combined / grid
futures.strategy	rsi / ema / combined
futures.leverage	leverage multiplier

Risk warning: the shipped defaults are SL 50% / TP 20% because that's what was requested. That is a 2.5:1 risk-to-reward AGAINST you and a 50% SL is unreachable on 3x+ futures (you'd liquidate first). Adjust before live trading.

4. Test the connection (recommended before running)

python -m src.check_connection

This verifies:

• .env keys are loaded
• Spot + Futures testnet are reachable
• API keys have trade permission
• Configured symbols return prices
• Wallet balances (warns if zero — go fund the testnet)

Sample output when healthy:

[ OK ] .env loaded — all 4 API keys present
[ OK ] Spot server time OK (drift 12 ms)
[ OK ] Spot account reachable — canTrade=True
[ OK ] Spot price BTCUSDT = 67234.50
[ OK ] Futures wallet: total=15000 USDT  available=15000 USDT
[ OK ] All checks passed — bot is ready to run

5. Run

# both bots
python -m src.main

# spot only
python -m src.main --spot

# futures only
python -m src.main --futures

Stop with Ctrl+C — it finishes the current tick and exits cleanly.

6. Deploy on Fly.io (recommended free cloud)

Fly.io runs the Dockerfile as a 24/7 worker. API keys live in Fly's encrypted secrets — never committed to the repo.

# 1. Install flyctl
#    Windows: iwr https://fly.io/install.ps1 -useb | iex
#    Mac:     brew install flyctl
#    Linux:   curl -L https://fly.io/install.sh | sh

# 2. Login
flyctl auth login

# 3. Pick a unique app name and create the app (does NOT deploy yet)
#    Either edit `app = "..."` in fly.toml first, or:
flyctl launch --no-deploy --copy-config --name <your-unique-name>

# 4. Set secrets (replace with your testnet keys)
flyctl secrets set \
  BINANCE_SPOT_API_KEY=xxx \
  BINANCE_SPOT_API_SECRET=xxx \
  BINANCE_FUTURES_API_KEY=xxx \
  BINANCE_FUTURES_API_SECRET=xxx \
  USE_TESTNET=true

# 5. Deploy
flyctl deploy

# 6. Watch logs
flyctl logs

# Useful ops
flyctl status                  # is it running?
flyctl secrets list            # what secrets are set (values hidden)
flyctl scale count 1           # ensure exactly 1 VM
flyctl ssh console             # shell into the running VM
flyctl apps destroy <name>     # tear it down

Updating config / strategy: edit config.yaml, commit, then flyctl deploy again. The new image rolls out with the new config.

Cost: the shared-cpu-1x / 256mb VM in fly.toml is the cheapest size. Fly used to include this in a free allowance; current pricing is roughly $2/mo. Confirm at https://fly.io/docs/about/pricing/

7. Docker (local or other cloud)

docker compose build
docker compose up -d              # both bots
docker compose logs -f spot       # tail logs
docker compose down               # stop

To run only one service: docker compose up -d spot or docker compose up -d futures.

Logs persist to ./logs/trades.log (rotated at 10 MB, kept 30 days).

Strategy notes

• RSI — buy when RSI crosses up through oversold, sell when it crosses down through overbought. Signals are evaluated on the most recently closed candle.
• EMA — buy on fast EMA crossing above slow, sell on cross below.
• Combined — both RSI and EMA must agree on the same direction; otherwise HOLD.
• Grid (spot only) — places levels limit BUYs below and levels limit SELLs above the current price, spaced by step_pct. Re-runs cancel and re-seed; restart the bot to refresh.

Going live

1. Run on testnet for at least a few days. Watch logs/trades.log and verify entries/exits make sense.
2. Set SL and TP to sane values (e.g., 0.02 / 0.05).
3. Lower order_size_usdt to a small live amount.
4. Set USE_TESTNET=false in .env.
5. Run — the bot prompts you to type I UNDERSTAND before starting in live mode.

Never commit .env.