Trust & security
Enterprise security by design
AgentCompliant is built from the ground up with security-first architecture for enterprises that can't afford to compromise on AI governance.
Security architecture
Defense in depth for governed AI
Encryption, identity, auditability, and isolation are first-class — not bolted on after the fact.
End-to-end encryption
AES-256 at rest, TLS 1.3 in transit.
Sensitive fields encrypted with industry-standard algorithms.
Role-based access control
Clerk-powered RBAC with org-level isolation.
Least-privilege defaults for every user and service account.
Immutable audit trail
Append-only audit tables with hash chain verification.
Tamper-evident history across engines and gateway events.
API authentication
Bearer token + API key dual auth on all endpoints.
Scoped keys with plan-aware rate limits at the gateway.
Data isolation
Per-organization schema isolation.
No cross-tenant leakage — every query scoped by org.
Infrastructure
OCI cloud with VPC isolation.
Automated backups and monitored availability targets.
Compliance frameworks
Built for regulated industries
Hover a badge for the full framework name. Roadmap items are planned attestations and programs.
Data handling
Your data stays yours
- US-based OCI infrastructure — data residency aligned with our hosted regions and your contractual requirements.
- Configurable retention policies per organization for audit and compliance evidence.
- Full data deletion on account termination — including GDPR Article 17 right-to-erasure workflows.
- We never use your data to train AI models — your governance data is not repurposed for model training.
Sub-processors
- Clerk — authentication and organization management
- Stripe — billing and subscription management
- Anthropic — AI analysis for compliance and product features (under data processing terms)
Infrastructure
Reliability and protection
- 99.9% uptime SLA target for the hosted platform.
- Automated database backups with tested restore procedures.
- DDoS protection via Cloudflare at the edge.
- Dual-window token bucket rate limiting on all API endpoints.
Responsible AI
Governance for AI, governed by principles
Bias monitoring — continuous signals and evaluations to surface drift and unfair outcomes before they reach production scale.
Explainability — traceable decisions and documented reasoning paths so teams can answer “why” for auditors and executives.
Human-in-the-loop (HITL) — approvals and escalations where autonomy must yield to human judgment.
Kill switch — instant containment when risk exceeds policy, with audit evidence of who acted and when.
Ready to secure your AI agents?
Start a trial or talk to us about architecture reviews, DPA terms, and enterprise deployment options.