How access to ZooIQ content works
Three classification tiers, three ZooIQ roles in Zoo IAM, and what each visitor can see.
How access to ZooIQ content works
Every ZooIQ page carries a classification. Your Zoo ID roles decide which classifications you can open. Nothing is decided by e-mail address or by guessing: the roles come from Zoo IAM at sign-in.
Classifications
| Classification | Who can read it | Typical content |
|---|---|---|
| Public | Everyone, no sign-in | Application overviews, user guides, this page, the API description |
| Internal | Signed-in users holding the ZooIQ reader role | Administration guides: where an application lives, how it is operated, which commands exist. Never secrets. |
| Restricted | Users holding the ZooIQ restricted or admin role | Infrastructure audit, network layout, deployment runbooks |
A page without a classification is treated as Restricted. Search, the Ask assistant, the REST API and the MCP server apply the same filter: results never include a page above your tier, and the answer engine only reads pages you could open yourself.
Roles
ZooIQ registers one OpenID Connect client in Zoo IAM (zoo-iq) with three client roles:
reader— Internal tier. Intended as the default role for Zoo employees through their verified e-mail domain policy.restricted— Restricted tier. Granted individually.admin— Restricted tier plus the Administration page (content reload, statistics). Marked privileged in Zoo IAM, so it can never be granted by a domain policy, an access request or an organization Admin — only by a Zoo IAM administrator by hand.
Roles are read from the ID token and access token that Zoo IAM issues to ZooIQ (resource_access.zoo-iq.roles, or the zoo_roles claim). A signed-in user without any ZooIQ role sees exactly what an anonymous visitor sees, plus a Request access link that opens the Zoo IAM request form for this application.
Sessions
ZooIQ keeps sessions in memory for at most 12 hours. Signing out ends the ZooIQ session and then the Zoo IAM session. When you sign out elsewhere, Zoo IAM tells ZooIQ through back-channel logout and the ZooIQ session ends too. A ZooIQ restart signs everyone out; nothing else is lost because ZooIQ stores no user data.
Assistants and API clients
Clients of the REST API and the MCP server present a bearer credential: either a ZooIQ API token issued by an administrator (stored only as a hash, with a fixed tier) or a Zoo IAM access token obtained for the zoo-iq client (tier from the roles inside it). Without a credential they see Public content only.