Agentic
Market
Intelligence
132 intelligence plugins process every bar, layer by layer—from raw market microstructure to AI-driven regime shifts. Our native Command Center provides end-to-end traceability, illuminating the precise path your signals take from ingestion to final execution. We don’t just generate signals; we provide the evidence-based mastery you need to dominate the market’s underlying mechanics.
4 Layers of Intelligence
132 plugins · I1 → I8 · each layer builds on the layer belowLayer 01
Data Foundation
Ingestion · Bar Building · Stream Distribution
- Institutional real-time tick + bar ingestion
- 1m → 5m / 15m / 1h / 4h / 1d aggregation
- Redpanda stream distribution
- 24 instruments · 6 timeframes
< 10ms pipeline latency · feed-provider bound, not processing bound
Layer 02
Mathematical Intelligence
58 plugins
- I1: RSI, MACD, ATR, VWAP, OBV, Supertrend, Bollinger Bands, CVD, OFI (28)
- I2: MACD/RSI/stoch events, exhaustion score, momentum accel (10)
- I3: Swing H/L, S/R, anchored VWAP, Fibonacci zones, market profile (8)
- I4: GARCH, Kalman, HMM, session context, VIX regime, cross-asset (12)
RSI 67.4 · MACD bullish crossover · GARCH: vol elevated · regime: trend
Layer 03
Pattern Intelligence
74 plugins + 2 aggregators
- I5: RSI/CMF/MACD divergence, squeeze, chart patterns (16)
- I6: BOS/CHoCH, FVG, order blocks, killzones, AMD cycles, CTF confluence (32)
- I7: 36 setup plugins: trend, mean-rev, SMC, session extremes
- CIS: 6-bucket convergence gate · score ±0.35 · 3/6 buckets required
BOS confirmed · unfilled FVG 5235–5238 · CIS +0.71 · CHoCHReversal fired
Layer 04
AI Intelligence
AI agent swarm · Ollama inference
- 4-agent swarm: skeptic, correlation, regime_coherence, counterfactual
- Local Ollama inference (gemma4:e4b default · configurable via .env)
- Full audit trail: per-agent latency, parse rate, token burn, outcome scoring
- Cross-asset narrative synthesis (6 groups, 1m–1h timeframes)
"Bullish 5m ES setup: trend + SMC confluence. FVG entry 5236, target 5258, stop 5229."
Platform
How It Works
Agentic Swarm Architecture
Asynchronous Intelligence
Our swarm of agents functions as a unified, high-performance intelligence engine. By decoupling every component through a durable, replayable event bus, the system scales effortlessly, allowing our intelligence capabilities to evolve in lockstep with the market’s complexity.
Validated Alpha Lifecycle
8-Class Hypothesis Tracking
Every signal is treated as a data-proven hypothesis. We track performance from the moment of activation through to final resolution, recording granular MAE/MFE metrics against an 8-class outcome model to turn every market move into a labeled training sample.
Regime-Aware Predictive Modeling
Multi-Model Consensus
Our system dynamically adapts to the current market environment by synthesizing GARCH volatility modeling, Kalman filters, HMM state detection, and BOCPD changepoint identification. This provides a multi-dimensional view of market regime, ensuring our intelligence is always contextually relevant.
Self-Optimizing Alpha Loop
Evidence-Gated Feedback
The loop from outcome to optimization is closed. System performance multipliers are automatically adjusted based on verified PnL and Sharpe ratio roll-ups, allowing our agentic intelligence to self-sharpen its edge without requiring manual code changes.
Predictive AI Narrative
Multi-Agent Swarm Synthesis
Beyond technical triggers, our I8 intelligence layer runs a swarm of specialized AI agents—skeptic, correlation, regime_coherence, and counterfactual—each contributing a scored multiplier to the signal. Full audit trail records per-agent latency, parse rate, and outcome to drive adaptive routing.
Ground-Truth Feature Store
Universal Intelligence Vector
Our feature store is the backbone of our Alpha generation. It maintains a persistent record of every instrument's state across every timeframe and tier, creating a labeled, high-resolution dataset that ensures every output acts as a training sample for the next iteration.
Architecture
Engineering Principles
Dependency-Ordered Execution
DAG · Topological sort · No polling
Every plugin declares its inputs. The engine runs Kahn's topological sort at startup. Execution order is a mathematical property of the dependency graph, not a convention anyone maintains. Cycles hard-crash at startup. Silent corruption is impossible. Adding a plugin means declaring its dependencies; ordering is inferred.
Regimented Signal Validation
CIS · 6-bucket convergence gate · Evidence required
When 5–8 setup plugins fire on the same bar, CIS decides what gets published and whether anything does. Six buckets (Trend, Momentum, Structure, Pattern, Institutional, Regime) must converge: score > ±0.35 AND 3 of 6 must agree on direction. One dominant bucket cannot override the rest. Discipline enforced by the architecture, not by policy.
Data as Ground Truth
Feature store · Outcome tracing · Self-improving
Every signal is tagged with the weight version that produced it. Every outcome (stop hit, target reached, TTL expired) is written back to the feature store with the full signal vector. Nothing is dropped. When outcome data is sufficient, weights update from evidence. The pipeline captures what it produces and learns from what it captured.
Canonical Typed Vocabulary
IntelligenceEvent schema · Stream-native · API-first
Every output at every tier is encoded into a versioned IntelligenceEvent schema and published to the stream bus. Producers publish. Consumers subscribe. No service calls another directly. A new consumer (alert engine, execution system, ML scorer) subscribes to the existing stream without changing the producers. The bus is the API. Extension is additive.
Live Trading Signals
Real-time signals across futures, forex, and crypto markets