Security

How ProteinInsight protects your account, isolates your data, and hardens the platform.

Trust isn't a feature — it's the foundation

Rigor, reproducibility and isolation come as standard.

Research use only

Every prediction-bearing screen carries the research-use disclaimer — never for clinical diagnosis, prescription, or treatment.

Reproducible evidence

Every result is tagged with its source, computation type and confidence, so findings can be verified — not taken on faith.

Multi-tenant isolation

Data is scoped per organization with role-based access control and audit logging on sensitive actions.

Technical posture

The concrete controls in place today.

Authentication & sessions

Password login issues short-lived access tokens; long-lived refresh tokens are minted separately and cannot be used to call protected APIs. The signing secret is required at boot — the service fails closed without it.

Multi-tenant isolation

Every program, candidate, job and report is scoped to your organization, and every query is filtered by that tenant key so one account can never read another's data.

Transport & browser hardening

TLS everywhere with HSTS, a strict Content-Security-Policy, and no third-party scripts — the 3D engine and web fonts are self-hosted. Framing and MIME-sniffing are blocked.

Data handling & privacy

Your sequences and analyses are never used to train models. External lookups are restricted to an allow-list of public scientific databases.

Platform administration

Platform-wide controls (e.g. legal/consent documents) are gated to a dedicated platform-admin allow-list, separate from — and not grantable through — ordinary tenant roles.

Security is continuously reviewed. Responsible-disclosure reports are welcome via your account contact.