The only data infrastructure platform that makes compliant medical image collection the default - built for MedTech companies training AI.
Three converging regulations are creating an existential risk for companies that haven't acted yet.
Raw Data exports can carry patient identity across their metadata tags. Storing or sharing these without pseudonymization is a GDPR (DSGVO) Art. 9 violation - right now, today.
GDPR (DSGVO) Art. 9 · Active violationA signed form cannot answer "did patient X consent to AI training vs. research vs. third-party sharing?" without manual review. AI Act Art. 10 demands structured, queryable provenance.
EU AI Act Art. 10Notified bodies require annotator qualifications, IRR metrics, and versioned dataset snapshots. Without it, your conformity assessment fails - regardless of model performance.
MDR Annex IV · Benannte StelleThree steps. One platform. No compliance rebuild later.
Patient signs granular consent on a tablet kiosk - separately for storage, research, AI training, and third-party sharing. Structured, version-locked, and cascade-able on withdrawal.
Meta data identity tags stripped automatically. Three isolated databases - medical data, identity data, and a cryptographic linker - make re-identification structurally hard.
R&D engineers download datasets with AI Act Art. 10 provenance reports and MDR Annex IV documentation built in. Every download logged. Every invalidation notified.
Built for the specific intersection of GDPR (DSGVO), EU AI Act, and MDR - not adapted from a generic SaaS tool.
Separate consent scopes for storage, research, AI training, future models, and third-party sharing. Timestamped, version-locked, and structured - not a PDF.
Three physically isolated databases. A breach of any single database reveals nothing. The split-identity pattern is the GDPR (DSGVO) pseudonymization requirement in infrastructure form.
Every upload, download, consent change, and access event logged with actor, timestamp, and reason. Exportable audit reports for supervisory authorities and notified bodies on demand.
One-click technical documentation export for notified body submissions. Ground truth metadata, annotator qualifications, IRR metrics, and versioned dataset snapshots included.
A patient withdraws consent → studies automatically removed from all active datasets → all engineers who downloaded are notified within 24h. Deletion logged and confirmed.
Ingest annotations from any tool via API. Pseudonymized annotator profiles, IRR metrics, and versioned snapshots stored alongside your images.
Every study collected without proper compliance documentation today is a problem tomorrow. Start building your compliant dataset now.
GDPR (DSGVO)-konform · EU data residency · Made in Germany