Scans US-listed stocks trading at $5 or less (NASDAQ / NYSE / AMEX only — no OTC) across premarket (7:00–9:30), regular (9:30–16:00), and after-hours (16:00–20:00) ET, and sends Telegram signals with exact buy and sell entry prices that already account for slippage, spread, and SEC/FINRA fees. Targets 3–5% net profit per trade. Optional auto-execution via Robinhood or Webull (off by default).
pip install -r requirements.txt
python main.py
Until Telegram is configured, signals print to the console.
- In Telegram, message @BotFather →
/newbot→ pick a name. Copy the bot token. - Message your new bot anything (e.g., "hi").
- Open
https://api.telegram.org/bot<TOKEN>/getUpdatesin a browser and copy the"chat":{"id": ...}number. - Put both values in
config.yamlundertelegram:.
🟢 BUY SNDL [premarket] [SIGNAL ONLY]
Setup: VWAP dip — 1.4% under VWAP, RSI 27
Last: $1.9400
BUY ENTRY (limit): $1.9449
SELL ENTRY (limit): $2.0163 (+3.0% net)
Stretch target: $2.0552 (+5.0% net)
Stop loss: $1.8896
Size: 51 sh (~$99)
Est. round-trip costs: 0.67% (already baked into targets)
The sell entry is computed so that after modeled slippage (0.25%/side), spread, and SEC + FINRA TAF sell fees, your net gain is at least the 3% floor. If costs would eat the edge (wide spread, illiquid name), no signal is sent.
Long-only, two setups:
VWAP mean-reversion — price ≥1% below session VWAP with RSI ≤ 30. Volume breakout — price clears the 30-bar consolidation high on ≥2x average volume.
Liquidity filters: ≥$500k daily dollar volume, estimated spread ≤1%, warrants/units/rights excluded. All tunable in config.yaml.
execution.auto_trade is false by default, and dry_run is true. To place real orders you must flip both — deliberately.
Know the risks: Robinhood and Webull have no official retail trading API. This app uses community libraries (robin_stocks, webull) that automate the same endpoints their apps use. That violates their terms of service and accounts have been restricted for it. Use at your own risk, and consider keeping only your trading capital at that broker.
Steps: uncomment the broker lines in requirements.txt, pip install them, fill credentials in config.yaml, run python brokers.py login-rh once for Robinhood MFA, then set auto_trade: true (keep dry_run: true for a few sessions first). All orders are limit orders — never market.
Max 3 open positions; max 3 day trades per rolling 5 days (PDT rule — with under $25k on a margin account, a 4th day trade flags your account; a cash account avoids PDT but T+1 settlement ties up funds); 90-min per-ticker cooldown; per-trade stop loss (~2.5% net); signals suppressed when the target exceeds 1.5x the day's demonstrated range.
- No strategy guarantees 3–5% per trade — that's the target the math is built around, not a promise. Expect losers; the stop loss caps them.
- Free data (yfinance) is delayed ~1–2 min in extended hours and thin premarket bars can be noisy. Validate in signal-only mode for a couple of weeks before trusting it with auto-execution.
- Extended-hours fills on Robinhood/Webull require limit orders (enforced here) and spreads are wider — expect fewer premarket/after-hours signals; that's the filters working.
- SEC fee rate (
sec_fee_per_million) changes periodically — verify the current rate at sec.gov and updateconfig.yaml. - Sub-$5 stocks are volatile and occasionally halted. Position sizing ($25–100) keeps any single blowup small.
main.py loop/scheduler · scanner.py universe · strategy.py signals & cost math · notify.py Telegram · brokers.py execution · config.yaml all settings · signals.csv / trades.csv logs (created on first run)