Security is a practice, not a promise
Protecting a workspace requires carefully configured infrastructure, verified access boundaries and ongoing operational review. These pages describe our security design, not an independent certification or a claim that attacks are impossible.
Access is explicit
The administrative gateway requires verified multi-factor authentication, an approved account and the correct organisation. Sessions expire after inactivity and have an absolute lifetime. Permissions are checked on the server, not inferred from a hidden button.
Only deliver what is needed
Private API routes and response fields must be explicitly approved. Response sizes, result counts and request volumes are bounded. Bulk export is not enabled by default, and private responses are excluded from public caching.
Intelligence without unchecked authority
Optional AI-assisted review works on aggregated security signals rather than raw customer records. Its output is a recommendation for review; it cannot grant access, publish private data or rewrite firewall rules.
Operations complete the picture
Deployment requires verified identity settings, protected upstream services, shared session storage, edge controls, retained audit events and tested backup restoration. No interface can prevent every copy of information already shown to an authorised person.