Kakunin โ AI Agent Compliance Demo
Rogue AI agent caught and cert-revoked in real time
AI agent compliance, cryptographic identity, EU AI Act, MiCA, X.509, behavioral monitoring, agentic frameworks
Issue X.509 certificates, monitor behavioral risk in real time, and generate auditable compliance reports โ all via a single API.
One API, three layers of trust โ from cryptographic identity to behavioral audit trails.
Issue X.509 certificates backed by AWS KMS RSA-2048 keys. Private key material never leaves KMS โ only the ARN is stored.
Stream behavioral events, score risk in real time, and trigger automatic certificate revocation when the score crosses the 0.85 threshold.
Generate MiCA Art. 70-72 and EU AI Act compliance reports on demand, backed by an immutable WORM audit log.
Register, certify, and monitor โ all REST. No SDKs required.
import requests BASE = "https://kakunin.ai/api/v1" HDRS = {"Authorization": "Bearer kak_live_...", "Content-Type": "application/json"} # 1 โ Register agent + issue X.509 certificate agent = requests.post(f"{BASE}/agents", headers=HDRS, json={ "name": "trading-bot-v2", "model": "gpt-4o", "version": "2.0.0", }).json() agent_id = agent["data"]["id"] cert_pem = agent["data"]["certificate"]["pem"] # RSA-2048 ยท 365 days # 2 โ Send a behavioral event result = requests.post(f"{BASE}/events", headers=HDRS, json={ "agentId": agent_id, "actionType": "transaction_initiated", "details": {"amount_usd": 500, "venue": "NYSE"}, }).json() risk_score = result["data"]["risk_score"] # 0.30 โ medium revoked = result["data"]["revocation_check_queued"] # 3 โ Stream live behavioral events via SSE import sseclient stream = sseclient.SSEClient(f"{BASE}/agents/{agent_id}/events/stream", headers=HDRS) for event in stream.events(): print(event.data)
Every design decision traces back to a regulation.
Hands-on demos โ no account needed for sandbox mode.
Everything you need to integrate Kakunin in your stack.