Skip to main content

Users

Represent the identity or account to which persistent information belongs. User scope lets information remain available across separate conversations.

Sessions

Represent a particular conversation, workflow, or period of activity. Helps maintain short-term continuity without treating every detail as permanent.

Namespaces

Represent a higher-level isolation boundary, such as an organization, workspace, or customer account, when supported by the deployment.

​
Users and sessions serve different purposes

  • Use a stable application-level identifier for each user.
  • Create or resolve a session identifier for each conversation or workflow.
  • Associate requests with the correct tenant or namespace when multi-tenant isolation is required.

​
Example identity model

​
Data isolation

  • Authenticate every request.
  • Authorize access to the requested user, session, and tenant.
  • Validate scope identifiers on the server.
  • Prevent one tenant from querying another tenant’s information.
  • Apply consistent access rules to retrieval, updates, and deletion.
  • Avoid logging secrets or unnecessary personal information.
  • Test cross-user and cross-tenant access boundaries.

​
Session lifecycle

A typical application can create or resolve a session when a conversation begins, attach relevant events to that session, retrieve session context during the interaction, and apply a retention or cleanup policy when the session ends.

​
Privacy and retention

Give users appropriate control over their stored information. Define retention periods, deletion behavior, access policies, and any process for exporting or correcting data where required.

​
Testing checklist