Deterministic, explainable, and auditable by design
FastPII focuses on deterministic sensitive data detection, clear transformation steps, and deployment choices that let organisations understand and review what their AI security layer is doing.
Operational trust
Security leaders can inspect how FastPII identifies, transforms, routes, and records AI traffic without relying on a black-box model judgement layer.
Core security properties of the platform
FastPII keeps the security story straightforward: deterministic detection, protected credentials, controlled transport, and auditable operations.
Deterministic detection
The detection layer does not rely on black-box ML. It uses explainable logic and checksum validation where supported.
Encrypted credentials
Provider credentials are encrypted at rest with Fernet using AES-128.
TLS in transit
Traffic between components is designed to move over TLS when communicating with providers and services.
BYOK
Keep provider access tied to your own key management decisions rather than a shared unmanaged path.
Self-hosting and EU-hostable
Run FastPII in self-managed environments and place deployments in EU-hostable infrastructure when needed.
Audit trails
Record request decisions and exports so operations and security teams can review what happened after the fact.
Security controls that stay legible as you scale
FastPII is built so platform teams can reason about the path from prompt to provider and from provider response to audit evidence without hidden layers.
Whether you start with the SDK, a managed API, or the AI Gateway, the same model applies: inspect deterministically, protect before forwarding, keep provider credentials protected, and retain the event trail for later review.
Explore the platform, then take the deeper security path
Start in the playground, inspect the docs, and bring FastPII into the AI request path that matters most to your organisation.