Sistava

Security Practices

Platform Security Overview

Cloud Security

Production runs in hardened cloud infrastructure with limited administrative access, private service networking, and clear ownership for every operational layer. Infrastructure is managed as code so changes are reviewed, repeatable, and auditable instead of improvised directly on live systems.

Encryption Everywhere

All customer data is encrypted in transit with TLS 1.3 and at rest with AES-256. Encryption keys are managed through dedicated key management infrastructure and rotated on a regular schedule. Internal communication between services runs over the cluster network with strict policy enforcement. No customer data ever traverses the public internet unencrypted.

Authentication and Access Control

User authentication uses industry-standard JWT cookies over HTTPS only, with passwords hashed via Argon2 and PBKDF2. Single sign-on via Google and Microsoft is supported for organizations that need it. Two-factor authentication is enforced for all administrative access. Role-based access control governs every action inside the workspace. The principle of least privilege applies to people, AI employees, and integrated tools alike. Every privileged action writes an audit log entry tied to the responsible identity.

Container and Network Security

The platform runs on a hardened Kubernetes environment with default-deny network policies, pod security standards, and ingress rate limiting. Container images are built from minimal base images and scanned for vulnerabilities before deployment. The application layer applies strict security headers, content security policy, and parameterized database queries throughout. Dependencies are continuously scanned for known vulnerabilities and patched on a documented timeline.

Tenant Isolation

Customer environments are isolated at every layer of the stack: database, application, cache, file storage, and workflow engine. Personal data, files, and AI conversations never cross tenant boundaries. Tenant identity is enforced by middleware on every request and by integration tests that probe for cross-tenant access. This is the most important security boundary on the platform and we treat any regression as a critical incident.

Data Security and Retention

Connected-app reads are minimized to the task at hand and are not bulk-copied by default. Raw third-party content is only kept as long as a live task needs it, and long-lived storage is limited to the specific product systems that need it. Retention is explicit and lifecycle-based. Operational records, memory, files, and logs each have a defined purpose, and customer deletion requests are handled through documented data-removal processes.

AI Safety and Desktop Control

AI employees operate inside explicit tool, task, and permission boundaries. Guardrails, approvals, and scoped access keep high-risk actions visible and under control. Desktop control is local-first and revocable. When a user starts the companion app, the cloud system acts as the brain while the local machine provides the hands, and users can pause or stop it at any time.

Monitoring and Recovery

Audit logs, runtime tracing, and operational monitoring provide visibility into what the system is doing and who triggered it. We use that visibility to investigate anomalies and respond quickly. Backups, recovery procedures, and incident response are treated as core security controls, not add-ons. We test recovery paths so the system can be restored cleanly after a failure.

Compliance and Assurance

We align the program with privacy and compliance obligations including GDPR, CCPA, and the EU AI Act. Those commitments shape how we collect, retain, disclose, and protect data. Security reviews, patching, dependency hygiene, and least-privilege access are part of the operating rhythm of the company.

What this means for customers