Skip to main content
Create and revoke API keys for your Cube organization. Admins can also perform these actions over the Platform API.

Personal API keys

Only admins can hold API keys unless you turn on Allow all users to create API keys on the Admin → Settings page. With it on, every user reaches Admin → API Keys and manages their own keys there — an admin still sees every key in the account, while everyone else sees only theirs. A personal key carries exactly its owner’s permissions, so it never reaches data or settings its owner can’t reach in the app, and it can only be scoped to deployments its owner has access to. Those permissions are what keeps a personal key narrower than an admin’s, so the switch can only be turned on while your account enforces roles and permissions. Turning it back off doesn’t delete existing personal keys, but it stops them authenticating: every request made with one is rejected with 403 Forbidden until it’s turned on again. Endpoints that act on behalf of someone else still require an admin’s key — generating an embed session is the main one. An admin’s key can also get an MCP server token that acts as another user.

Deployment scope

By default, an API key is not scoped to any deployment—it can be used with every deployment in your organization (the All deployments option). When creating or editing a key, you can instead scope it to one or more specific deployments. The scope determines which requests a key can authorize:
  • Deployment-scoped endpoints—such as generating an embed session for a deployment—require the key to be scoped either to that deployment or to all deployments. A request targeting a deployment outside the key’s scope is rejected.
  • Endpoints that aren’t tied to a specific deployment work with any key, regardless of its deployment scope. The exceptions are endpoints that act on the whole organization rather than a single deployment. They require a key scoped to all deployments and reject any other key with a 403 error:
Each key’s scope—All or the specific deployments it’s limited to—is shown in the Deployments column of the API keys table.
Scoping is optional and backward compatible. Existing keys remain unscoped (All deployments) until you change them.

Endpoint scope

By default, an API key works with every endpoint (the All endpoints option). When creating or editing a key, you can instead limit it to specific endpoints, picked from a list grouped by API area. Selecting a whole group also covers the endpoints added to it later. A request to an endpoint outside the key’s scope is rejected with 403 Forbidden. The endpoint scope and the deployment scope apply together, so a request must pass both. Endpoints that return another credential, such as getting an MCP server token, reject every key limited to specific endpoints. Each key’s endpoint scope is shown in the Endpoints column of the API keys table. Over the Platform API, set the scope with allowedEndpoints: a list of endpoint IDs, such as AuditLogPublicController.exportAuditLog, or of whole groups, such as AuditLogPublicController.*. An empty list makes the key work with every endpoint.
Give an integration a key limited to the endpoints it calls. For example, a SIEM tool that pulls the audit log needs only AuditLogPublicController.exportAuditLog.