An autonomous multi-asset research and execution-governance platform. AI proposes, deterministic systems dispose.
STRATA is not a trading bot. It is a research and execution-governance platform that runs autonomously across four asset classes — crypto perpetuals, equity and commodity perpetuals, real equities, and FX. The system is built around a single architectural principle: the AI generates proposals; deterministic logic decides what actually executes.
The reasoning engine is REI Core 0.5a, developed by REI Labs — a hypergraph traversal engine, not a large language model. Ecliptica integrates REI Core as STRATA's reasoning layer and builds the multi-asset product and deterministic governance around it. Ten governance gates — Portfolio Manager, Entry State Machine, Risk Brake, Profit Ladder, Price Sentinel, and five more — sit below every AI decision and can approve, reduce, or reject regardless of AI conviction.
This documentation serves three audiences. Choose your path.
Why STRATA is differentiated, defensible, and structurally advantaged. Competitive positioning, moat analysis, what compounds.
The eight-layer stack, trading cycle flow, REI teaching infrastructure, and how each unit reasons. Technical detail without marketing.
Risk controls, safety nets, operational rigor, regulatory posture, and failure modes. Assumes NDA. Assumes professional counterparty.
For complete architecture detail, see architecture. For the full unit inventory and teaching-infrastructure reference, see units.
STRATA runs an eight-layer stack. Data flows top-to-bottom; governance flows bottom-to-top. Every layer is independently observable and independently governed.
A full architecture reference, including the trading-cycle state machine and the teaching loop, is available in the architecture section.
REI Core evolves traversal parameters per unit via a genetic algorithm. Two units starting from the same config diverge over time into domain-specialized reasoners.
Patterns, principles, and relationships become inference nodes in the hypergraph. Reinforced knowledge strengthens; unreinforced knowledge fades.
Posture gate, exposure limits, entry-state machine, profit ladder, risk brake, re-entry controller, material-change controller, price sentinel, exposure graph, post-trade forensics.
Every detected mistake generates a correction fed back into the relevant unit. Auto-corrections fire for invariant violations. Manual corrections adjust reasoning pathways.
Fund audiences should continue to the edge. Operators should continue to how it works. LPs should continue to diligence.
STRATA is in selective deployment with institutional counterparties. For allocator diligence, partnership inquiries, or licensing discussions, contact Ecliptica directly.