Skip to main content

Customer Security FAQ

Status

Owner: Tech lead · Status: draft · Canonical source: Enterprise Customer Inquiries

Answers to common security, architectural, and data privacy questions raised by prospective enterprise customers.

1. Is our proprietary data used to train AI models?​

No, never. Flezi Galaxy utilizes commercial enterprise agreements with Azure OpenAI Service and Anthropic API. Both providers legally guarantee that customer data submitted via their enterprise APIs is never used to train or refine foundational models.

2. Can malicious code escape from the execution sandbox?​

No. Code execution occurs inside unprivileged Docker containers on dedicated sandbox nodes. The containers operate without root privileges (no_new_privileges = true), enforce strict process and memory limits, and are destroyed once tasks settle.

3. Where is our data physically stored?​

All primary data stores — including PostgreSQL database servers, Redis instances, Blob storage accounts, and Key Vaults — are hosted within Microsoft Azure's Southeast Asia (Singapore) region.

4. Can we export and delete all data upon contract termination?​

Yes. Organization administrators have the authority to request a full data export and trigger complete, permanent deletion across all database tables and cloud object storage containers.

Existing Controls​

  • Self-serve deletion tools accessible to Organization Administrators.
  • Documented data protection guarantees in enterprise MSAs.

Evidence (Files & Configuration)​

Limitations & Roadmap​

  • Dedicated Private Cloud or on-premises deployment options for regulated banking clients.