Skip to main content
Evaluates a guardrails policy by key. Send a JSON body with request.input as a string (the text to check, e.g. model output or user message). The API returns an envelope with meta (code, message, errors) and result (violated, decision, reason, key). Response shape
  • meta.code — HTTP status code
  • meta.message — Status message (e.g. Success, Policy evaluation failed)
  • meta.errors — Array of validation or evaluation error messages when present
  • result.violated — Whether the policy was violated
  • result.decision — ALLOW, BLOCK, or ESCALATE
  • result.reason — Human-readable reason (optional)
  • result.key — Policy key that was evaluated
Example response 
{
  "meta": { "code": 200, "message": "Success", "errors": [] },
  "result": {
    "violated": false,
    "decision": "ALLOW",
    "reason": "Content passed guardrails",
    "key": "content-guardrail"
  }
}