Skip to main content

Privacy & Security

How we protect your data and ensure secure inference.

Data Handling​

Your data is yours. We do not use your prompts, completions, or any data sent through the API to train models.

Request Privacy​

  • No training on your data — Inputs and outputs are never used to improve or fine-tune models
  • No logging of content — Request/response bodies are not persisted beyond operational needs
  • Ephemeral processing — Data exists in memory only during inference

Data Residency​

All inference runs on infrastructure located in the European Union. Data does not leave EU boundaries during processing.

Infrastructure Security​

Network Security​

  • All API traffic encrypted with TLS 1.3
  • DDoS protection at edge
  • Private network isolation between tenants

Compute Isolation​

  • Dedicated GPU memory per request
  • Memory cleared between inference calls
  • No shared state between users

Access Controls​

  • API keys scoped to specific workspaces
  • Audit logging for all API access
  • Role-based access in AI Studio

ISO Certificates & Other Audits​

For a up-to-date overview of our current certifications visit our Drata page: https://app.drata.com/trust/a89d9d62-a65a-4338-91eb-725517defa0c

Reporting Security Issues​

If you discover a security vulnerability, please report it to security@nebul.com.

Questions?​

Contact our team through the Support Portal for security-related inquiries.

Last updated on