Mnemos AI
Security

Security your security team designed.

Permission-aware retrieval, row-level tenant isolation, customer-managed keys, and a complete audit lineage on every AI response. Built so your CISO can say yes.

SOC 2 Type II
HIPAA-ready
GDPR / CCPA
ISO 27001 (in progress)
Customer-managed keys
EU residency
BYO LLM
SSO & SCIM

Row-level tenant isolation

Every tenant-scoped table uses Postgres row-level security. Our runtime database role cannot bypass it. The tenant context is enforced at the connection layer, not at the application layer.

Permission-aware retrieval

ACLs inherit from your source systems and Mnemos roles. Filtering happens before the model sees a candidate, not after. A restricted document is not just hidden — it is never retrieved.

Customer-managed keys

Bring your own KMS — AWS, GCP, or Azure. Encrypt the memory graph at rest with keys you control. Revoke access with one click.

Regional residency

US-East, US-West, and EU-West regions available. Enterprise customers can pin tenants to a region and block cross-region replication entirely.

Responsible disclosure

Found a vulnerability? Tell us.

Email security@mnemos.ai with details. We acknowledge reports within 24 hours and coordinate disclosure with affected customers.

Talk to a Mnemos security engineer.

We send the SOC 2 report, the DPA, the architecture diagram, and the sub-processor list before the first call.