12 AI-Powered Modules
Every tool a quantitative trader needs — from strategy generation to risk management, data sourcing to portfolio composition.
AI Strategy Builders
Five distinct AI modes for generating, predicting, and managing trades with large language models.
Regime Detector & Entry Builder
Classify market conditions into Trend, Range, Consolidation, Oscillation, or Bespoke regimes, then generate regime-aware entry signals. Two-layer approach separates classification from entry logic.
Kronos AI Predictor
Time-series forecasting with three model sizes (mini/small/base). Multi-dimensional signal filtering separates prediction accuracy from noise for robust entry signals.
Trader AI Entry Agent
LLM evaluates every price bar with full context — philosophy, indicators, and market state — for granular buy/sell/hold decisions on each candle.
AI Libero Interval Agent
LLM decisions at configurable intervals (every N bars) instead of every bar. 10-20x reduction in API calls while maintaining AI reasoning for swing trading.
AI Strategy Studio
Describe strategies in natural language, iterate via chat, get executable Python code. Eliminates the programming barrier for traders with strong intuition.
Risk Management & Filtering
Defense-in-depth exit rules and market condition filters to protect capital.
Market Observer Filter
Block trades during unfavorable conditions using indicator-based observation rules. Separates "should I trade?" from "what trade?" to reduce false signals and drawdowns.
Indicator Exit & Risk Manager
Five-rule exit system: Circuit Breaker, Time Limit, Regime Detection, Drawdown Limit, and Indicator Guard. Multiple independent safety layers prevent catastrophic losses.
Data & Visualization
Multi-source market data with local caching and real-time backtest analytics.
Multi-Source Market Data
Six providers: YFinance, Dukascopy, ClickHouse, Alpaca, CCXT, BaoStock. Local Parquet caching with zero-copy executor access and concurrent download queue.
Real-Time Results Analytics
Live equity curves, trade logs, and 7-metric dashboard (PnL, Sharpe, MaxDD, Win Rate, Profit Factor) streaming as backtest executes. Discover failing strategies early.
C++ Backtest Engine
500-1000x faster than Python. C++ executor with embedded Python via pybind11 delivers zero-copy NumPy access through Apache Arrow. Enables 100+ backtests per hour.
Free Deep Dive →Signal Composition
Institutional-grade multi-signal fusion with statistical weighting and factor screening.
Quant Lab Alpha Factory
Institutional-grade signal fusion combining 1000+ signals with 5 weighting methods: Equal, Confidence, Voting, Max Confidence, Min Confidence. IC/ICIR/Sharpe screening.
Начните свою фабрику сигналов
Уровень бесплатного пользования включает механизм бэктестинга C++, обнаружение режима и данные YFinance + Dukascopy — все, что вам нужно, чтобы начать масштабное строительство.