Platform Security
Traffic is designed to run over encrypted channels, and platform boundaries are segmented by workspace and project scope.
Built for teams that need observability, control, and predictable security operations.
AI Cost Board follows a shared-responsibility model. We secure platform infrastructure, routing, and access boundaries while customers secure prompt content, end-user data, and application-side authorization.
Traffic is designed to run over encrypted channels, and platform boundaries are segmented by workspace and project scope.
Role and membership boundaries are used to limit which users can view projects, provider integrations, and request logs.
Provider credentials are managed in project context and should be rotated regularly. Teams should follow least-privilege principles for all API keys.
Operational telemetry supports anomaly detection for error spikes, unusual spend behavior, and degraded provider performance.
Security checks are embedded into product development and release workflows, with priority handling for vulnerabilities and production issues.
Do not send unnecessary personal data in prompts. Use project-level isolation for staging and production. Review logs and budgets regularly.