ModelCopLLC ← Back to site
Platform · Architecture

Architecture

How ModelCop is built: a multi-tenant control plane with isolation enforced at the database kernel, encryption everywhere, and a small, auditable attack surface.

Platform

Tenant isolation (three independent layers)

Data protection

Monitoring & detection

Management-plane audit logging (CloudTrail), network flow logs, database logs, and application logs are captured to dedicated, access-controlled stores. Threat detection (GuardDuty) is enabled across the environment, and container images are scanned for vulnerabilities on every build.

Inline governance

ModelCop can sit inline between an agent and its LLM provider: point the provider base URL at ModelCop, and prompts are classified and checked against policy before they are forwarded — with a forensic record on every decision.

Disaster recovery

Recovery is validated end-to-end, not merely configured. A point-in-time restore exercise — new cluster, schema and role verification, tenant-isolation check, and teardown — has been completed and is re-validated periodically.

For a deeper walkthrough, request our Security & Compliance Overview at security@modelcop.ai.