How It Works

The trading engine runs a continuous 60-second cycle during US market hours. Here's what happens each cycle.

1

Universe Scanning

Each trading day, the system rebuilds its universe by identifying the top US equities ranked by average daily dollar volume, supplemented by major ETFs like SPY, QQQ, and IWM. Only stocks with sufficient price and liquidity pass the filters.

Ensures tight spreads, deep liquidity, and stable microstructure.

2

Market Data & Indicators

Every 60 seconds, the engine fetches 1-minute bar data for the entire universe and computes a suite of technical indicators: VWAP, RSI, ATR, volume ratios, range compression, and more.

All indicators are computed locally from raw bar data — no black boxes.

3

Strategy Signals

Four strategies run in parallel each cycle, looking for specific setups: VWAP mean reversion, opening range breakouts, trend-day momentum, and volatility breakouts after range compression.

Each strategy has precise entry and exit conditions tuned for its market pattern.

4

Regime Classification

A regime classifier evaluates each symbol's current market behavior — is it a mean reversion day, a trend day, an opening drive, or range compression? Signals are scored based on how well the strategy fits the detected regime.

Misaligned strategy/regime pairs are filtered out entirely. Only high-confidence, aligned signals pass through.

5

Risk Management

Surviving signals are validated against portfolio constraints: position sizing limits, maximum open positions, sector exposure caps, total portfolio exposure, and daily loss thresholds.

The highest-scored signals get capital priority. Risk limits are never bypassed.

6

Execution & Protection

Approved signals become bracket orders with a marketable limit entry, a broker-side stop-loss, and a broker-side take-profit. Positions are protected even if the server goes offline.

Software-side exit logic (trailing stops, VWAP crosses, time limits) can close positions earlier — the bracket legs auto-cancel.

AI Analytics Layer

On top of the core trading engine, an AI analytics layer provides human-readable insights:

  • •Trade Narratives — 3-5 sentence explanations of why each trade was entered and exited
  • •Daily Briefings — end-of-day performance summaries with actionable observations
  • •Interactive Chat — ask questions about your trading day with full context awareness
  • •Recommendations — parameter and strategy suggestions backed by your real performance data
  • •Real-Time Observations — alerts during trading for regime shifts, signal clustering, win/loss streaks, and risk events
Explore all features →