Finite recurring missions
Fixed-delay missions bind explicit action identities, approvals, start times and expiry.
Give AI a governed path into your business context—and an executable world to explore what happens next.
Select a component to explore · Logical architecture · Illustrative flow
An approved client discovers a narrow tool surface. Identity and policy define the boundary; simulation produces a result with an execution record.
MCP is the interface. CortexSim brings together the business semantics, simulation environment and execution records that make the connection useful.
Seven purpose-built tools for discovery, isolated simulation and evidence. The gateway exposes a deliberate interface to a much broader platform.
MCP / toolscortexsim_preview_transactionPreview up to 10 operations against a synthetic lab checkpoint without persisting effects.
Seven implemented source profiles. Separate trust boundaries. Source credentials stay with the connector, outside the AI model.
postgres-viewBring relational business context through scoped views and reviewed mappings.
Behind the gateway, CortexSim services address the lifecycle of an approved operation—from a finite mission to reconciliation, compensation and protected recovery.
Fixed-delay missions bind explicit action identities, approvals, start times and expiry.
Retained execution records support history, restart behavior and traceable review.
An uncertain result has an explicit reconciliation path before further action.
Supported action profiles provide a governed compensation path with its own evidence.
Independent custody protects restore decisions against erased or invalidated state.
Implemented for internal use within supported contracts and synthetic profiles. These platform services are distinct from the seven exposed MCP tools. Customer integration qualification and production deployment remain deferred.
Map one approved AI client, one private source and one consequential decision into a focused CortexSim evaluation.