Built for trust. Designed with care.

When you speak to a Companion, the audio flows from your device to our cloud, which is hosted in a secure Azure environment. We process it and send it to AI providers (such as Google or Microsoft) for further processing. We then process their response, store it in our encrypted storage, and deliver the final result to the end user. All data is encrypted in transit and at rest.

Yes. We are SOC 2 Type II compliant as a company, audited annually in April/May. SOC 2 compliance governs our processes—how we handle data, access controls, confidentiality, incident response, disaster recovery, and change management. These company-wide standards apply to all our products and operations.

Never. Training models on user data—by ourselves or third parties—is something we never do. We also never share your data or sell it. Our partners are contractually required to follow the same standards.

Access to customer data is restricted to authorized personnel who need it for maintenance, debugging, and operations. We follow the principle of least privilege, and all access is logged and audited as part of our SOC 2 compliance. Audit logs are maintained for 6 months.

Our system is multi-tenant with logical separation controls. Customer data is isolated using access controls and separation mechanisms to ensure data from one organization cannot be accessed by another.

When signs of a security breach appear, we investigate immediately. If confirmed, we identify the scope and impact, then notify affected customers per contractual commitments. We provide ongoing updates throughout investigation and remediation, and conduct post-incident reviews to prevent recurrence.

Regional data residency is available for enterprise customers. Data may be sent to different regions for AI processing, but once we receive a response, that data is deleted from processing regions. Your data at rest—including files, transcripts, and call history—can be stored in your specified region. This is a premium capability; contact our enterprise team to discuss requirements and pricing.

Companions are built with multiple safety layers: AI disclaimers clearly identifying them as AI-based representations, safety reviews to limit harmful behavior, continuous refinement of prompts and safety rules, and tiered crisis intervention protocols for sensitive situations. We actively monitor and update our safety systems.

No. All digital twins are restricted to users 18 and older. Digital twins can only be created by the individual themselves through a live consent video and verification process. Uploaded photos are screened against a database of known public figures, and unauthorized attempts to create twins of others are blocked.

Yes. For enterprise customers, we provide annual SOC 2 Type II audit reports, penetration test reports from independent auditors, and complete responses to security questionnaires. Contact our team to request documentation for your security review.

All Companions are clearly identified as AI-based, not licensed professionals or guaranteed sources of truth. Users consent to this understanding at sign-up. Companions provide contextual AI disclosure when advice could be mistaken for professional services (medical, legal, financial). We encourage responsible use and provide disclaimers to reduce the risk of misuse.