Market evidence
Completed price history, authoritative structure and the session’s primary area of interest.
02 — TECHNOLOGY
AI reasoning.Independent checks.
Our technology brings research, structured AI assessment and explicit software validation into one traceable process. Each layer has a distinct responsibility.
Explore the framework
01 / INPUTS
Build on verifiable market data.
The strategy reads composite-mid FX candles and quotes, with weekly, daily and intraday history used for different parts of the assessment. These reference prices provide analytical context; they are not assumed to be executable broker prices.
Freshness and integrity checks run before analysis. Missing confirmation candles or invalid higher-timeframe data prevent an assessment from proceeding on unreliable evidence. The system does not ask AI to invent missing prices.
Economic evidence
Relevant official releases and current, sourced news, evaluated for timing and direct market impact.
02 / AI
Reason within a defined framework.
AI receives structured market context and applies the supported strategy criteria. It evaluates whether the evidence describes a completed setup, what remains missing and whether the proposed risk plan is coherent.
Its response follows a defined schema, making direction, confirmation, news context and the plan available for explicit checks. AI supports research and interpretation; it cannot bypass the validator or independently authorise execution.
03 / BOUNDARIES
Keep validation separate.
Deterministic software checks required fields, price ordering, allowed markets, session timing, authoritative trend direction and risk arithmetic. A plausible explanation from AI does not override a failed condition.
The analysis layer and execution boundary are separate. Broker execution, when deliberately enabled, requires additional quote, sizing and protective-order checks. Website visitors have no connection to that operating environment.
See the validation boundaries04 / TRACEABILITY
A record worth revisiting.
Decision records preserve the assessment, supporting context, rule results and simulated trade lifecycle. Scanner and system events help explain when analysis ran, why a candidate was blocked and what happened afterwards.
Review is based on recorded evidence. Rule completion describes satisfied conditions, not a forecast of winning probability. Changes are evaluated deliberately rather than presented as automatic self-improvement.
