Security
Security you can build on.
Aethera is engineered with zero-trust principles and defense in depth, so your work stays private, segregated, and encrypted at every layer.
We never use your content to train models. Your data is yours to keep, and yours to delete.
How we protect your data
Defense in depth, at every layer.
Encrypted everywhere
TLS 1.3 with HSTS for everything in transit, and AES-256 for everything at rest.
Identity and access
JWT-based sessions and OAuth 2.0 sign-in with Google and Microsoft, with role-based access enforced on the client, the server, and the API.
Your data is segregated
PostgreSQL Row-Level Security keeps every organization’s data isolated, and parameterized queries are used throughout.
Hardened APIs
Rate limiting, strict input validation, idempotency keys, and HMAC-signed webhooks guard every entry point.
Locked-down infrastructure
An isolated processing environment, strict firewall and IP allowlisting, and secrets injected at runtime, with zero credentials in our code.
Monitoring and audit trails
Real-time error tracking and immutable audit logs for sensitive actions, so changes stay traceable.
Compliance
Built to meet the standards.
We are not yet certified, but our controls were purpose-built to map directly to these frameworks, and we are actively working toward certification.
SOC 2
In progressControls for logical access, system monitoring, and change management align with the Trust Services Criteria.
ISO 27001
In progressOur security management foundation aligns with key Annex A controls for access, operations, and development.
GDPR
Built inPrivacy by design: IP anonymization and row-level data segregation are part of the architecture.
HIPAA
Safeguards in placeThe technical safeguards for access control, audit control, and transmission security are implemented.
Reporting a vulnerability
If you believe you have found a security issue, please email us so we can look into it quickly and responsibly.
help@aethera.aiStart with the Free plan.
No credit card required. Starter credits are included, so you can try the agent, the connectors and every model from your first prompt.
