Skip to main content

Overview

This section explains why the outputs can be trusted as run artifacts without overclaiming what the platform proves scientifically.

What this section covers​

  • Security controls and operational boundaries
  • Data handling and trust boundaries
  • Auditability and run provenance
  • Cryptographic sealing and verification
  • Retention, deletion, and self-hosted trust differences

Trust model in one sentence​

Glassbox Bio is designed to produce evidence-linked, machine-verifiable, cryptographically sealed run artifacts. That means the platform is built to make computational execution and provenance auditable. It does not claim to replace scientific review, wet-lab confirmation, or customer governance policy.

Why this matters​

Users need a clean distinction between:

  • Confidence that the run artifact is authentic and reproducible
  • Confidence that the underlying science is sufficient for a downstream decision

This section exists to keep those two claims separate.