24 technical rules from the recommender engine evaluate your portfolio every 15 min. When the score exceeds the BUY threshold, the robot executes with automatic stop-loss and take-profit. Risk-free while you learn: shadow mode (0 risk) → paper → live real.
Every 15 min the engine scans your portfolio and watchlist. Applies technical rules (SMA, RSI, MACD, ATR, volume), fundamentals (P/E, EPS, market cap), analyst consensus (Wall Street ratings) and insider trading. Each rule adds or subtracts points to a 0-100 score. If the score exceeds the configured BUY threshold and the ticker meets all guardrails, the robot emits a signal.
Shadow: only logs what it would have done. 0 risk, ideal to validate the engine 2-3 weeks before scaling.
Paper: trades simulated portfolio (internal Finanzas or Alpaca paper if you have broker connected).
Live: sends real orders to broker (only with Level 1 permission activated + Telegram 2FA). Switching modes is a dropdown; never any ambiguity about what's happening.
Hard limits are in Python, not editable config: max 5% capital per trade, max 10 concurrent positions, max 20 trades/day, mandatory stop-loss on every order, 5s cooldown between orders, automatic circuit-breaker at 5% daily drawdown. Neither a bug nor an attacker can bypass them. You can tighten them from the UI, never loosen beyond the hard-cap.
The /robot page shows win rate, expectancy, max drawdown, cumulative P/L, distribution by rule and comparison vs S&P 500. Real data from your trades (shadow/paper/live), not optimistic backtests. With 30+ trades you already have solid statistical reading to decide whether to scale or iterate rules.
Demo mode with sample data · email verification in 30s · no broker connection needed.
