For enterprise teams

/docs/persona-enterprise-team

Roles, audit trails, model governance, and budget controls for organizations that need to prove who did what, and keep it that way.

Once an AI gateway sits behind more than one team, the questions shift from "how do I make a request" to "who can approve a new model, who can see billing, and can we show an auditor exactly what changed and when."

Role-based access, not shared credentials

Every workspace member holds a role built from a fixed permission taxonomy — read/manage splits for members, keys, agents, billing, budgets, models, and policies, plus dedicated audit:read and security:read permissions for people who only need visibility. A member can never grant a capability they don't personally hold, including through a key they create — see API key scopes for how that ceiling is enforced on keys specifically.

An audit trail you can hand to security

Every workspace mutation — a policy change, a role change, a key action — is written to the workspace activity log, with the actor, the target, and a timestamp. This is what turns "we think that changed on the 12th" into a record you can pull up during an incident review or a compliance audit, without reconstructing events from memory or Slack history.

Model governance: request and approve

Larger organizations often want a gate between "a new model exists" and "anyone on the team can call it in production." Model access supports a request/approve flow — a member can request a model be enabled for the workspace, and someone holding model-approval permission signs off before it's usable. Pair this with a model allowlist so the set of callable models reflects an actual decision, not just whatever a provider shipped that week.

Budgets that scale past one team

Spend controls exist at the workspace, member, and key level (see Setting spend controls), which is what lets a platform team structure keys by project and give each one its own ceiling without needing every project's budget pre-negotiated against a single workspace-wide number.

What "Enterprise" adds on top

The plans above roles, audit, and governance — SSO and a dedicated support relationship — are part of Relixr's Enterprise tier and are set up directly with the team rather than self-served from the dashboard today. If your organization needs those, that's a conversation to have before rollout, not a settings toggle to hunt for afterward.

Tip

Roll out access control before you roll out the integration, not after. Define roles and turn on the model allowlist while the workspace has three users, not three hundred — retrofitting governance onto an already-sprawling set of shared keys is the harder direction.

Was this helpful?

Still stuck? Help center · Doctor