Security & Privacy

Yourdata,yourrules.

Enterprise AI adoption is limited by trust, not capability. Cortex makes security, auditability, and deployment boundaries visible from day one.

Data isolation

Every tenant operates in a fully isolated environment. Your data is never mixed with other organizations. Indexes, storage, and processing are scoped to your tenant.

Traceability

Every AI-generated output links to its source documents with inline citations. Full audit trails for all user actions, queries, and generated artifacts.

Access control

Container-based RBAC with role inheritance. Admins, editors, and viewers. Private containers block inheritance for sensitive content. Permissions enforced at the service layer.

Deployment flexibility

Cube5-hosted single-tenant, customer VPC with private networking, or on-premises deployment. Choose where your data lives based on your compliance requirements.

AI governance

AI should augment human judgment, not replace it. Cortex enforces guardrails at every stage of the AI pipeline.

AI outputs are advisory — human review gates before finalization

No customer data used for model training

LLM calls are logged and traceable via Langfuse

Citation-first design: claims without evidence are flagged

Model selection configurable per deployment

Input/output filtering for sensitive content

Compliance

Built to meet the highest security standards for enterprise and heavily regulated industries.

SOC 2 Type II (in progress)

GDPR-compliant data processing

Data residency options (EU, US, custom)

Right to erasure support

Data processing agreement available

Regular third-party security assessments

Questions about security?

We're happy to walk through our security architecture, share our security documentation, or discuss your specific compliance needs.

Talk to our security team