Getting Started with AGEI

Sign up and create your organization. AGEI uses tenant isolation to ensure your evidence vault remains separate from other organizations.

Create a policy set that defines governance rules for AI model deployments, agent actions, or data processing activities.

{
  "policy_name": "Model Deployment Review",
  "scope": ["model_deployment"],
  "rules": [
    {
      "condition": "model_risk_score > 0.7",
      "action": "require_human_approval"
    }
  ]
}

Link your policy to a gate that evaluates lifecycle events. Gates produce outcomes like approve, deny, escalate, or inspect.

Use the AGEI API or Python SDK to submit lifecycle events to your gate for governance evaluation.

POST /api/v1/gates/:gate_id/evaluate
{
  "event_type": "model_deployment_requested",
  "payload": {
    "model_id": "credit-risk-v3",
    "risk_score": 0.85
  }
}

Every evaluation creates a tamper-evident receipt. Query receipts for audit trails, compliance reporting, or incident investigation.

{
  "receipt_id": "rcpt_001",
  "outcome": "escalate",
  "reason_code": "HUMAN_REVIEW_REQUIRED",
  "content_hash": "sha256:a1b2c3...",
  "created_at": "2026-05-04T16:30:00Z"
}