Skip to main content

The governance API for AI

Evaluate any action against 7 mathematical governance laws in under 15ms. Get an instant verdict: PROCEED, ESCALATE, or BLOCK — with a tamper-proof evidence trail.

Get StartedAPI Reference
One API call to govern any AI action
curl -X POST https://api.arqera.io/api/v1/ara/governance/evaluate \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $ARQERA_API_KEY" \
  -d '{
    "action": "data.delete",
    "description": "Delete all inactive user accounts",
    "context": {
      "risk_level": "high",
      "is_irreversible": true,
      "has_approval": false
    }
  }'
Instant verdict with per-law results
{
  "verdict": "block",
  "action": "data.delete",
  "explanation": "High-risk irreversible action requires human approval",
  "duration_ms": 11.2,
  "evaluations": [
    { "law_id": "safety_dominance", "result": "fail",
      "reason": "High-risk irreversible action without approval" },
    { "law_id": "bounded_autonomy", "result": "warn",
      "reason": "Agent trust score below threshold for deletion" },
    { "law_id": "audit_conservation", "result": "pass", "reason": "..." },
    { "law_id": "budget_conservation", "result": "pass", "reason": "..." },
    { "law_id": "evidence_gravity", "result": "pass", "reason": "..." },
    { "law_id": "least_action_path", "result": "pass", "reason": "..." },
    { "law_id": "monotonic_truth", "result": "pass", "reason": "..." }
  ]
}
⚡

Quickstart

Govern your first AI action in 5 minutes. One API call, instant verdict.

📖

API Reference

Full documentation for the Governance Evaluate endpoint, Evidence API, and more.

🐍

Python SDK

pip install arqera — clean Python client with typed results and error handling.

📦

TypeScript SDK

npm install @arqera/sdk — full TypeScript types, Express middleware, Next.js examples.

🇪🇺

EU AI Act Compliance

Map ARQERA’s governance engine to EU AI Act requirements for Annex III systems.

⚖️

The 8 Laws

Understand the 8 mathematical governance laws that evaluate every action.