Pre-deployment EVF intake
Maps AI risk, reversibility, governance readiness, behavioural envelope, and failure-mode exposure.
Features
Platform architecture for runtime AI governance.
OBEXGATE is organised around six governance dimensions, operational dashboards, deployment controls, and integration paths.
Six dimensions
The six dimensions of OBEXGATE governance
Runtime policy engine
Blocks or holds unauthorised actions before execution.
Hash-chained evidence
Captures every action, including refused actions. The author cannot be the final verifier.
Named human owner
Every action carries a named human owner. Delegation does not erase accountability.
PRISM outputs
Renders the same evidence for governance, audit, regulatory, and executive audiences.
Jurisdiction controls
Treats jurisdiction, region, residency, and data sovereignty as first-class platform fields.
Dashboard and operations.
Operational surface
Coverage, risk, alerts, inventory, gap analysis, risk scoring, and regulatory timeline.
Live control feedback
Governance and approval events through WebSocket-SSE push.
Scenario simulation
Test enforcement, denial, drift, discovery, and approval events before operational release.
Continuity by design
Responsive metrics with collector fallback and cached dashboard state.
SDK and integration posture.
OBEXGATE is designed to govern the action boundary. It can sit beside existing systems, inside product workflows, or across federated deployments without depending on one model or cloud provider.