FHIR Pseudonymizer¶

Rule-based de-identification

Describe what should happen to each element with FHIRPath rules: redact, generalize, shift dates, hash, encrypt, remove, or replace with a pseudonym.

De-identification

Pseudonymization backends

Create and reverse pseudonyms with gPAS, Vfps, entici or any service implementing the MII pseudonymization interface.

Pseudonymization services

Per-request settings

Serve many projects from one deployment: override rule settings, or send an entire anonymization config along with the request.

Dynamic rules and config

Streaming from Kafka

Consume FHIR resources straight from Kafka topics and publish the result, with dead-lettering and at-least-once delivery.

Reading from Kafka

Production ready

Release images are signed with cosign and ship with SLSA Level 3 build provenance. They are built on a minimal, distroless base image and are executable following NSA/CISA runtime hardening practices. Prometheus metrics and OpenTelemetry tracing are built in.

Image verification

Observability

Fast

Tens of thousands of requests per second on a single instance, with full horizontal scalability. Low cold-start latencies allow for automatic scale to 0.

Performance