Security Model
Understand the trust principles and security boundaries built into Nuxari.
Last updated: June 2026
Who this is for
Admins, security teams, and procurement reviewers who want to understand the trust and security design of the Nuxari platform.
Before you start
This page covers the platform's trust principles in user-facing terms. For specific configuration questions, see Workspaces and Roles, Connectors, or Edge Agents.
Trust principles
Tenant isolation
Your organization's data is completely isolated from every other organization on the platform. Every query, evaluation, finding, evidence record, and audit event is scoped to your workspace. No data is shared, mixed, or accessible across organizations, ever.
Role-based permissions
Every action in Nuxari is checked against the requesting user's role before it runs. A user cannot view, modify, or initiate actions outside the scope of their assigned role. Roles are enforced at the platform level, they cannot be bypassed by a UI interaction or an API call.
Approval-controlled execution
High-impact actions, such as revoking access, offboarding a user, or rotating credentials, require human authorization from a designated approver before they execute. The platform will not proceed with any such action without a completed approval record.
Immutable audit trail
Every action taken in the platform, by a user, a scheduled job, or the system itself, produces an immutable audit event. Audit events cannot be modified or deleted after they are written. The audit log is a complete, time-ordered record of everything that happened in your workspace.
Credential handling
Integration credentials are stored securely and are never returned to browser clients or included in API responses. Connector credentials are only used server-side to authenticate against external systems. Credential age is tracked and surfaced as a finding when rotation is overdue.
AI guardrails
The AI Governance Assistant is a recommendation layer only. It cannot grant or revoke access, call external systems directly, bypass the approval workflow, or make final authorization decisions. All actions it recommends must pass through the standard approval process before execution.
Edge agent security
Edge Agents operate outbound-only, they initiate connections to Nuxari's cloud service and never accept inbound connections from outside your network. This means your network perimeter stays closed. Registration tokens are one-time-use and can be revoked at any time from your admin console.
Responsible disclosure
If you believe you have found a security issue in the Nuxari platform, please contact us atsecurity@nuxari.com. We take all security reports seriously and respond promptly.