bitget-sma-bot

Open-source Bitget trading bot in Python — pluggable strategies (SMA crossover, Bollinger breakout, RSI mean-revert) for USDT-M perpetual futures, with free demo testnet, paper trading, control dashboard, and Telegram alerts.

🇬🇧 English | 🇫🇷 Français · Installation · Documentation · Disclaimer

Why this project exists

Most crypto bots are either paid black boxes or a few-hundred-line scripts impossible to hack on. bitget-sma-bot is deliberately small (~2,000 lines of Python), readable in an evening, and meant to be a solid starting point for anyone building a systematic crypto trading workflow on Bitget USDT-M perpetuals:

✅ 3 pluggable strategies (SMA crossover, Bollinger breakout, RSI mean-revert)
✅ 4 execution modes: paper (simulated), dry (logs only), demo (Bitget virtual account), live (real money)
✅ Full control dashboard: start / stop the bot, tweak the strategy, watch live PnL — without touching a single line of code
✅ Built-in backtester + grid-search optimizer
✅ Telegram alerts, stop-loss, take-profit, % of balance risk sizing
✅ End-to-end tested on Bitget demo: real orders, position tracking, PnL verification

At a glance

Component	Details
Strategies	SMA crossover, Bollinger breakout, RSI mean-revert — all inherit `Strategy(ABC)`. Add yours = 1 file
Modes	`paper` (no API key), `dry` (logs), `demo` (real orders on Bitget's virtual account via `paptrading: 1` header), `live`
Risk	% of balance per trade, stop-loss and take-profit as % from entry, checked on every tick
Notifications	Telegram on every entry / exit / error (optional)
Backtests	`backtrader` engine (classic) + small engine for fast grid search
Tests	54 pytest tests + 9 end-to-end tests on Bitget demo + 6 UI tests (Playwright)

1-click quick start

Windows: double-click install.bat, edit .env, then start.bat (the bot) or start_web.bat (the dashboard).

Linux / macOS:

chmod +x *.sh
./install.sh
./start_web.sh    # dashboard on :5000

Open http://localhost:5000 and click Start Bot.

The control dashboard

Everything happens here. No code editing required.

Status hero: RUNNING / PAUSED / STOPPED badge, uptime, Start / Stop / Pause / Restart buttons
5 stat cards: balance, position, unrealized PnL, last price, last signal
TradingView chart: candles + SMA fast / slow + entry line. Independent symbol + timeframe picker (look at ETH while the bot trades BTC)
Equity curve rebuilt from past trades
Activity log: bot stdout live-tailed every 4 seconds
3 config panels: Strategy (switch SMA / Bollinger / RSI + dynamic params), Risk (RPCT, SL, TP), Mode & Alerts (paper / dry / demo / live + Telegram)
Force close button: closes the current position on the next tick (uses reduceOnly exchange-side)

Architecture

See docs/architecture.md for the long form.

Backtest results

Conditions: 180 days of real Bitget data, 1,000 USDT starting balance, 2% risk per trade, SL 3%, TP 6%, 0.06% commission. Reproducible:

python examples/showcase.py --days 180

Best config per pair and timeframe

Symbol	TF	Strategy	Trades	Win %	Return	Max DD	Sharpe
ETH	1d	Bollinger 20/2	9	66.7%	+0.56%	0.40%	+1.38
ETH	4h	RSI 14 30/70	52	57.7%	+0.85%	0.44%	+0.63
ETH	1h	SMA 20/50	41	43.9%	+0.51%	0.63%	+0.24
BTC	4h	RSI 14 30/70	49	59.2%	+0.47%	0.35%	+0.43
BTC	1h	SMA 20/50	35	42.9%	+0.32%	0.46%	+0.19
BTC	1d	Bollinger 20/2	7	42.9%	+0.20%	0.34%	+0.66

Full grid in examples/results/showcase/summary.csv.

Capital preservation through a 30% crash

Over the same window, BTC dropped from ~91k to ~63k (-30%). Buy-and-hold ended around 700 USDT. The three strategies held the line near 1,000 USDT.

The bot doesn't ride bull markets. It sidesteps big drawdowns. That's the canonical trade-off of a systematic system.

Available strategies

Name	Module	Params	Idea
`sma_crossover`	src/strategies/sma_crossover.py	`fast`, `slow`	Long when fast SMA crosses above slow, short on the reverse.
`bollinger`	src/strategies/bollinger.py	`period`, `std`	Long on breakout above upper band, short on break below lower band.
`rsi_mean_revert`	src/strategies/rsi_mean_revert.py	`period`, `oversold`, `overbought`	Long when RSI rebounds from oversold, short when it falls from overbought.

Adding your own = create a file in src/strategies/, inherit Strategy, register it in __init__.py::STRATEGIES. The trader and the optimizer pick it up automatically.

# src/strategies/my_strategy.py
from .base import Strategy, Decision

class MyStrategy(Strategy):
    name = "my_strategy"
    def __init__(self, lookback=14):
        self.lookback = lookback
    def warmup(self) -> int:
        return self.lookback + 1
    def signal(self, df) -> Decision:
        # your rule here
        return Decision("flat", "not yet implemented")

Bitget demo mode (recommended before going live)

Bitget offers a demo account with virtual money. The bot supports it through the paptrading: 1 header (see Bitget official docs).

🎁 No Bitget account yet? Sign up via the project's referral link (code 9K5D7K4J).

What you get: the current Bitget welcome bonus + fee discounts (the exact amount is set by Bitget and shown on the signup page). What the project gets: a share of the trading fees you pay to Bitget (no extra cost on your side). This is how the bot stays free and open-source — full disclosure. For the affiliation to count: sign up via this link (the code is pre-filled), complete KYC. Commissions only trigger on real trading volume, not on demo mode.

https://www.bitget.com/asset/demo-trading — activate the demo account
Switch to demo mode in the Bitget dashboard (top of the page)
Personal Center → API Key Management → Create Demo API Key (separate from live keys)
Permissions: Read + Trade + Futures. No Withdrawal.

Drop the values in .env:

BITGET_API_KEY=...
BITGET_API_SECRET=...
BITGET_API_PASSWORD=...
MODE=demo

Sanity-check everything before running the bot:
```
python scripts/test_demo_connection.py
```

Tests

pytest                                        # 54 tests, ~2 min
python scripts/e2e_full_demo.py               # 9 end-to-end tests on Bitget demo
python scripts/test_ui_smoke.py               # 6 Playwright UI tests

Coverage:

Unit: indicators, strategies, trader (paper / dry / demo / live), SL / TP, registry
API contract: every Flask endpoint + input validation
E2E exchange: real market orders on Bitget demo, position checking, PnL accounting, edge cases
UI smoke: Playwright headless, Start / Stop click flow, dropdowns, JS error guard

Tech stack

Layer	Choice
Language	Python 3.11+
Exchange connection	ccxt (multi-exchange, configured for Bitget)
Backtests	backtrader + lightweight in-house engine
Notifications	python-telegram-bot v20+
Dashboard	Flask + Tailwind CSS (CDN) + lightweight-charts
Tests	pytest + Playwright
Containerization	Dockerfile + docker-compose

Project layout

src/
  config.py              .env loading + mode / strategy validation
  exchange.py            ccxt Bitget wrapper (sets the paptrading header in demo mode)
  indicators.py          SMA, EMA, RSI, Bollinger bands
  strategies/            pluggable framework
    base.py              Strategy(ABC), Decision, Signal
    sma_crossover.py     SMA strategy
    bollinger.py         Bollinger strategy
    rsi_mean_revert.py   RSI strategy
    __init__.py          registry + make_strategy() factory
  trader.py              main loop: SL/TP check, signal, paper/demo/live dispatch
  backtest.py            backtrader engine
  simple_backtest.py     small bar-by-bar engine for grid search
  telegram_notifier.py   async Telegram client
  web.py                 Flask dashboard
  main.py                bot entry point

examples/
  run_backtest.py        backtrader-based backtest
  optimize.py            grid-search optimizer
  showcase.py            18 backtests + comparison chart

scripts/
  test_demo_connection.py   Bitget demo smoke test
  e2e_full_demo.py          9-test E2E suite
  test_ui_smoke.py          Playwright UI tests
  take_screenshots.py       dashboard capture

tests/                  54 pytest tests

docs/                   architecture, strategy, setup

Disclaimer

Educational tool. Do not use live mode with money you cannot afford to lose. Trading futures with leverage can wipe out a balance in minutes. Read DISCLAIMER.md before going further.

License

MIT.

If this project helped you, drop a ⭐ — it helps other devs find it.

Name		Name	Last commit message	Last commit date
Latest commit History 26 Commits
docs		docs
examples		examples
scripts		scripts
src		src
tests		tests
.env.example		.env.example
.gitignore		.gitignore
DISCLAIMER.fr.md		DISCLAIMER.fr.md
DISCLAIMER.md		DISCLAIMER.md
Dockerfile		Dockerfile
INSTALL.fr.md		INSTALL.fr.md
INSTALL.md		INSTALL.md
LICENSE		LICENSE
README.fr.md		README.fr.md
README.md		README.md
docker-compose.yml		docker-compose.yml
install.bat		install.bat
install.sh		install.sh
pyproject.toml		pyproject.toml
requirements.txt		requirements.txt
start.bat		start.bat
start.sh		start.sh
start_web.bat		start_web.bat
start_web.sh		start_web.sh
stop.bat		stop.bat
stop.sh		stop.sh

Navigation Menu

Repository files navigation

bitget-sma-bot

Why this project exists

At a glance

1-click quick start

The control dashboard

Architecture

Backtest results

Best config per pair and timeframe

Capital preservation through a 30% crash

Available strategies

Bitget demo mode (recommended before going live)

Tests

Tech stack

Project layout

Disclaimer

License

About

Releases

Packages

Contributors

Languages

Folders and files

Latest commit

History

Repository files navigation

bitget-sma-bot

Why this project exists

At a glance

1-click quick start

The control dashboard

Architecture

Backtest results

Best config per pair and timeframe

Capital preservation through a 30% crash

Available strategies

Bitget demo mode (recommended before going live)

Tests

Tech stack

Project layout

Disclaimer

License

About

Topics

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages