Features
CHIMERA / FULL FEATURE SUITEGBM Direction, HMM Regime, CascadePredictor, DeepLOB, and RL Agent analyze independently and vote as the AI Council. A separate 4-source confirmation gate (Technical / Sentiment / OnChain / DeepLOB cross-check) must also agree before execution.
VPIN, cascade risk, funding divergence, OI, CVD exhaustion, whale activity, sentiment, model disagreement, and more stress every position before entry.
6-stage validation: data collection, feature engineering, hyperopt, walk-forward, Monte Carlo, deploy gate. No parameter ships without statistical proof.
Drawdown circuit breakers, consecutive loss limits, volatility gates, leverage caps, BTC veto, whale veto. Graduated cooldown: 2h to 24h.
21 specialized modes across alpha, passive, scalping, arbitrage, and operator-controlled workflows. Strategy Mux routes mode families by HMM regime state.
View Bot ModesVWAP, TWAP, Almgren-Chriss, Iceberg, Smart Limit, Market. Volume-scaled slippage model, spread simulation, funding rate accounting.
Hidden Markov Model classifies market state in real time: trending, mean-reverting, volatile, transitional. Hurst exponent confirms persistence.
Technical indicators, order flow, funding rates, open interest, volume profile, whale movements, sentiment analysis, on-chain metrics, cross-exchange divergence.
Signal approval is only half the job. Ghost evaluates market impact, fill quality, liquidity, and volatility before routing, delaying, splitting, reducing, or vetoing execution.
Exchange to backend to Redis to Socket.IO to UI. Every supported runtime metric streams in real time. No polling for product telemetry.
Slippage, spreads, funding, liquidation, Monte Carlo, and Deflated Sharpe — plus a calendar-day breakdown that exposes the losing streaks every cumulative equity curve hides. A 1y +50% result with 60 consecutive red days isn't hidden anymore.
Combinatorial Purged Cross-Validation prevents look-ahead bias. Walk-forward windows validate parameter stability across regime changes.
Most trading bots stop at signal generation. ChimeraMind goes further: it reads market sentiment, measures stress regimes, checks model consensus, applies risk gates, and uses Ghost Protocol to decide how, or whether, a trade should reach the market.
ChimeraMind combines social activity, market news, AI-powered text sentiment, and fear/greed context into a single composite market psychology score. Instead of reacting to raw hype, the system checks source diversity, sentiment divergence, volume spikes, and confidence before exposing insights to dashboards or sentiment-aware bots.
Most bots focus on when to trade. ChimeraMind also focuses on how the trade should be executed. Ghost Protocol sits after signal generation and before order placement, helping reduce poor fills, reckless entries, and avoidable market impact.
Ghost does not expose internal infrastructure on the public site. It presents the execution logic users need to understand: impact control, adaptive routing, fill-quality feedback, and trade vetoes when execution quality is poor.
Ghost Protocol is not a profit engine. It is execution intelligence: the layer that tries to make a valid signal enter the market with better timing, sizing, routing, and discipline.
Antifragile measures market disorder across volatility, liquidity, cascade risk, funding pressure, sentiment conflict, and model disagreement. The system reduces exposure first, then uses capped allocation to search for asymmetric stress opportunities when risk gates allow it.
Start with paper. Move to live when you're ready.
Sign up free — no card required. Connect your exchange in the desktop app.