stratXRequest a briefing

HOW STRATX WORKS / TECHNICAL DEPTH

One decision. Independent authority. A recorded result.

This page explains the technical model behind the product views. StratX structures a proposed action, evaluates participant-owned constraints, narrows the available paths and preserves the basis for review.

TARGET MODEL / FINANCE PROTOTYPE SCOPE MARKED

SYSTEM BOUNDARY

Participants remain independent.

Asset Manager, broker, funding, custody and clearing roles retain their own policies, approvals and actions. StratX coordinates the shared decision state.

Neutral layerSTRATXCOORDINATIONNO EXECUTION

Select a participant to inspect its authority. Each remains outside StratX.

FROM INTENT TO EVIDENCE

The sequence behind the decision.

Canonical Intent Object

A versioned representation binds the scoped proposed action and its provenance.

Policy evaluation

Applicable governing and institution-owned constraints are evaluated under each participant’s authority.

Feasibility Set and RouteScore

Required gates define permissible paths. Only those paths may be compared; a score never grants permission.

Coordination and evidence

The result, blockers, participant responses and handoff are recorded for authorized inspection.

TECHNICAL REFERENCECORE COMPONENTS

The model, in detail.

Canonical Intent Object (CIO)The proposed action

A versioned, canonical representation of an intent and its domain context. Adapters supply scoped inputs from existing systems. Accepted inputs and the full evaluation context must be bound before reproducibility can be claimed.

Participant Policy PacksIndependent constraints

Applicable governing requirements and internal institutional policies remain distinct. Each participant owns and approves the policies represented under its authority. Policy evaluation cannot create that institution’s approval.

Feasibility SetPermission before preference

The bounded set of candidates that satisfy the required participant assertions and hard gates in the evaluated context. Blocked, conditional, waiting or unknown requirements cannot silently enter the permissible set.

RouteScoreEligible comparison

Ranks permissible alternatives under a bounded scoring contract. A score cannot override policy, replace approval, or make an unavailable candidate permissible. Selection and execution remain institutional actions.

Decision TraceThe recorded determination

Preserves relevant input provenance, policy context, participant responses, route exclusions and the resulting determination. Later authorization or action events are separate records, with their own provenance.

Decision ReconstructionWhat was known at the time

Retrieves the retained context for a historical determination. Re-evaluation under changed policy is a new result. Exact executable replay additionally requires the retained evaluator artifacts and separate verification.

RIMAuthorized regulator inspection

A role-scoped audit view for authorized regulators and auditors. Inspection does not grant institutional authority or initiate external actions. Production regulator access is not established in the current evidence.

Institutional handoffThe execution boundary

StratX evaluates, coordinates, explains and records. Institutions retain selection, authorization and execution through their own systems. No assets, private keys or institutional execution authority move into the coordination layer.

GOVERNANCE & FAILURE

The boundary holds when a dependency does not.

Governance gates precede comparison. Missing required responses remain unresolved, and fail-closed behavior prevents an unknown from appearing permissible.

StratX returns

Permissible alternatives, blockers, reasons and Decision Trace. Decision Reconstruction retrieves retained context as of a recorded time; exact executable replay requires separate verification.

Institutions act

Participants select, authorize, execute and settle. RIM is a role-scoped, read-only model for authorized oversight; it gives no execution authority.

IMPLEMENTATION SCOPE

What the prototype demonstrates.

Synthetic multi-participant finance coordination, path evaluation, ranking snapshots and scoped audit views. Live institutional integrations, production regulator access, historical evaluator retention and SAFE_MODE enforcement need separate verification.

THE NEXT CONVERSATION

Discuss a technical evaluation.

Start with one decision, institution-owned policy scope and the evidence needed to test the model.

Request an institutional briefing