REGISTERED SYSTEMS
—
— in production
ENTERPRISE CONTROL PLANE
Checking API
GOVERNANCE POSTURE
Good morning. Your AI estate is under control.
A single operational view of risk, controls, evaluations, and evidence across every AI system.
—
POSTURE
EVALUATION COVERAGE
—
Production gate readiness
HIGH + CRITICAL RISK
—
Require enhanced controls
AUDIT EVENTS
—
Immutable evidence records
SYSTEM INVENTORY
Risk distribution
RELEASE READINESS
Governance gates
LIVE ACTIVITY
Recent decisions
AI SYSTEM REGISTRY
Systems under governance
Ownership, deployment, risk, and control status in one authoritative inventory.
CONTROLLED RUNTIME
Gateway lab
Test model routing and enforcement through the same governed path used in production.
Request Parameters
Governance Security Result
OPENAI COMPATIBLEThe response, detector verdicts, groundedness, and enforcement action will appear here.
POLICY ENGINE
Effective controls
Risk-tier defaults resolved against system-specific governance configuration.
RELEASE ASSURANCE
Evaluation gates
Submit detector coverage evidence before promoting AI systems to production.
Submit coverage
Release gate status
Choose a registered system to inspect its latest status.
IMMUTABLE EVIDENCE
Audit trail
Structured decision records for investigations, assurance, and regulatory review.
DELIVERY QUEUE
Notifications
What the platform told people, and whether it arrived. A message that never lands is a workflow waiting on someone who was never asked.
ADVERSARIAL TESTING
Red team
Run the attack corpus against a system's own resolved controls. Nothing reaches a model provider — these are attacks, and having one answered proves nothing that inspecting it does not.
Choose a system and run a campaign.
USING gardai labs
Help
What each part of the control plane is for, and the guarantees it makes.
HTTP API
API reference
Generated from this server, so it describes what is actually running.