Crypto Trader

A small, hackable crypto trading bot with a web dashboard. Describe a strategy, drop in one Python file, run it, and watch positions and P&L live.

• Backend: FastAPI, SQLAlchemy, SQLite (Postgres optional), pure-Python indicators, httpx for market data.
• Frontend: Vue 3 + Vite + Tailwind dashboard with a live WebSocket feed.
• Brokers: paper (simulated fills, default) and binance (live spot).
• Market data: Binance public REST (no API key needed).
• Backtesting: replay historical candles through the same strategy interface, with return/drawdown/win-rate/Sharpe metrics — so what you test is what runs.

Paper trading is the default. Live mode places real orders and is opt-in. This is experimental software. Never trade more than you can afford to lose.

How it works

Binance REST ──▶ MarketData ──▶ TradingEngine ──▶ Broker (paper | binance)
                                     │
                           Strategy.on_candle(ctx)
                                     │
                     REST /api + WebSocket /api/ws ──▶ Vue dashboard

The engine runs one strategy at a time. Each tick it fetches candles, updates the mark price, enforces stop-loss/take-profit, calls the strategy, executes the resulting intents, and persists orders, equity snapshots and events.

Quickstart (local)

# 1. Config
cp .env.example .env

# 2. Backend
python3 -m venv .venv
.venv/bin/pip install -r backend/requirements-dev.txt
cd backend && ../.venv/bin/python -m uvicorn app.main:app --reload --port 8000

# 3. Frontend (new terminal)
cd frontend && npm install && npm run dev

Open http://localhost:5173. API docs are at http://localhost:8000/docs.

Quickstart (Docker, local)

cp .env.example .env
docker compose -f docker-compose.local.yml up --build
# dashboard: http://localhost:8080   API docs: http://localhost:8000/docs

The default docker-compose.yml is the deployment stack (Traefik + docker_default); use docker-compose.local.yml for a self-contained local run.

Deployment

Merging to main triggers .gitea/workflows/deploy.yml, which tests, builds both images, and rolls the stack out on the shared Traefik host:

• The frontend joins the external docker_default network and is published at crypto.${BASE_DOMAIN} (default crypto.matitos.org) via Traefik, with a Let's Encrypt cert (certresolver=myresolvercd).
• The backend stays on the internal network only; the frontend nginx proxies /api (including the /api/ws WebSocket) to it, so API keys are never directly exposed.
• The WebUI is protected by fail-closed Basic auth: the deploy aborts unless CRYPTO_FRONTEND_AUTH_USERS is set (falling back to the shared SCA_FRONTEND_AUTH_USERS). Generate a value with htpasswd -nbB user 'password'.
• A new release is rolled back automatically if healthchecks fail.

Repo/org variables and secrets used by the deploy:

Name	Kind	Purpose
CRYPTO_FRONTEND_AUTH_USERS	var/secret	htpasswd string for the WebUI (required)
BASE_DOMAIN	secret	Base domain (default matitos.org)
CRYPTO_MODE	var	paper (default) or live
CRYPTO_EXCHANGE / CRYPTO_SYMBOL / CRYPTO_TIMEFRAME	var	Trading defaults
CRYPTO_STARTING_BALANCE / CRYPTO_POLL_INTERVAL	var	Paper account + tick rate
CRYPTO_EXCHANGE_API_KEY / CRYPTO_EXCHANGE_API_SECRET	secret	Required only for live
CRYPTO_EXCHANGE_SANDBOX	var	Use the Binance testnet
CRYPTO_CORS_ORIGINS	var	Defaults to https://crypto.<BASE_DOMAIN>
CRYPTO_DATA_DIR / PERSISTENT_MOUNT	var	Host path for the SQLite DB
DOCKERHUB_USERNAME / DOCKERHUB_TOKEN	secret	Optional; for authenticated pulls

Configuration

All settings are environment variables (see .env.example):

Variable	Default	Purpose
MODE	paper	paper or live
EXCHANGE	binance	Market data / execution venue
SYMBOL	BTC/USDT	Default trading pair
TIMEFRAME	1m	Candle interval
POLL_INTERVAL	5	Seconds between ticks
STARTING_BALANCE	10000	Paper account size
FEE_BPS / SLIPPAGE_BPS	10 / 5	Paper fill model
DATABASE_URL	sqlite:///./data/trader.db	Postgres: postgresql+psycopg://...
EXCHANGE_API_KEY / EXCHANGE_API_SECRET	—	Required only for live
EXCHANGE_SANDBOX	false	Use Binance testnet

Adding a strategy

Drop a file in backend/app/engine/strategies/ — it is auto-discovered and appears in the UI immediately. See AGENTS.md for the full template and the StrategyContext API. Two examples ship in the repo: sma_cross and rsi_reversion.

API

Method	Path	Description
GET	/api/status	Bot status, equity, P&L
GET	/api/strategies	Available strategies + parameter specs
POST	/api/bot/start	Start a strategy
POST	/api/bot/stop	Stop the bot
POST	/api/backtest	Backtest a strategy on historical candles
GET	/api/positions	Positions (with unrealized P&L when open)
GET	/api/orders	Order history
GET	/api/events	Activity log
GET	/api/equity	Equity curve points
POST	/api/orders	Manual market order
POST	/api/positions/{id}/close	Close the open position
WS	/api/ws	Live status + event stream

Tests & lint

make test     # backend pytest + frontend vitest
make lint     # ruff
make build    # frontend production build

Project layout

backend/
  app/
    engine/            trading loop, indicators, strategies/
    broker/            paper + binance execution
    market/            binance + fake data feeds
    api/routes.py      REST + WebSocket
    models.py          positions, orders, equity, events
    store.py           persistence
  tests/
frontend/
  src/                 Vue dashboard