Trust center v5

Compliance posture for distributed AI infrastructure.

UmamiEdge needs to earn trust from site owners, enterprises, regulators, and AI customers. v5 introduces the operating controls needed for an investor-ready pilot.

Security model Audit log Safe SQL docs

Core controls

What v5 adds for production conversations.

Control

Tenant isolation

Organization-scoped data model, RLS policies, service-role-only writes for sensitive infrastructure workflows, and selected-tenant context.

Control

Node identity

Per-node API keys, SHA-256 hashed secrets, key prefix lookup, revocation workflow, and future signed telemetry support.

Control

Operational audit

Runtime registrations, incident state changes, key lifecycle operations, and admin actions are captured in audit_events.

Control

Responsible AI

OpenAI-compatible gateway stubs are designed for model allowlists, usage limits, abuse monitoring, and local data residency rules.

Control

Infrastructure safety

Power headroom, temperature, latency, and critical telemetry can generate incidents before routing capacity is expanded.

Control

Commercial readiness

Usage events, GPU seconds, estimated costs, and billing account tables support pilot economics and partner revenue share.

Policy pages to finalize before launch