Security

The platform is multi-tenant with logical isolation at organisation and brand level. Brand DNA, assets, campaign history and agent configuration are scoped to their owning tenant and are never used to serve another tenant’s request.

• 01All traffic is served over TLS. Plain HTTP is not accepted.
• 02Data at rest is encrypted by the managed storage and database layer.
• 03Credentials are stored as salted hashes. Plaintext passwords are never stored or logged.

• 01Invite-only account creation. There is no open self-service signup.
• 02Role-based access control with least-privilege defaults.
• 03Session-based authentication with server-side validation on every privileged request.
• 04Administrative actions are logged with the acting user and timestamp.

Platform integration credentials are held in a server-side secret store and are never exposed to the browser. Integration tokens are scoped to the minimum permissions required and can be revoked by the client at any time from the integrations panel.

• 01Retrieved brand context is treated as data, not as instructions, to reduce prompt-injection risk from third-party content.
• 02Agent runs execute against scoped tool permissions; an agent cannot reach a resource outside its brand scope.
• 03Budget and spend-related outputs are advisory until a human approves them; no autonomous spend is executed by an agent.

The platform runs on managed cloud infrastructure with automated backups of the primary datastore. Restore procedures are exercised as part of platform maintenance.

Report suspected vulnerabilities to [email protected]. Please include reproduction steps and avoid accessing or modifying data that is not yours. We acknowledge reports and will keep the reporter informed through remediation.