Built to be trusted with regulatory work
How ReGentra constrains model outputs, protects your data, and preserves auditability across every medical device assessment.
Why generic AI isn't enough and
why data handling can't be an afterthought
A general purpose model can produce plausible sounding technical files, but its claims often lack traceable links to regulation. That taught us to build the other way round: verify regulation and evidence first, then constrain every generation to what the sources support.
ReGentra's principle is to structure what can be structured, automate what scales, and keep a human reviewer in control.
ReGentra applies data controls aligned with EU requirements: EU hosting, minimal collection, explicit retention windows, tenant isolation, strong authentication, and full audit logging, all designed in from the start.
Three boundaries inside the Simulation module
Every Simulation output passes through the same trust model.
Why the system is built this way,
and how your data is protected
Each engineering decision addresses a specific way AI systems fail in regulated contexts.
Each data control governs where information lives, who can reach it, and how long it stays.
Closed evidence boundary
The model cannot access external knowledge or generate claims beyond what the evidence base contains. This isn't a prompt instruction — it's an architectural constraint that holds regardless of what the model attempts.
Human approval required
Every output remains a draft until a reviewer explicitly approves it. There is no override, no batch-approve, and no silent promotion. The human is the final authority at every decision point.
Traceability & verification
Generated outputs are verified against their source material and every claim is linked to the source it came from. Anything that cannot be traced back is treated as an error and flagged for reviewer action.
Full audit trail
Every generation, edit, approval, and rejection is recorded. The trail exists so the process can be demonstrated and defended — who reviewed what, when, and what they decided.
Specified before built
The system's rules, structures, and boundaries were defined by regulatory domain expertise before implementation began. The specification came first; the code was built to meet it.
Defensive by default
The system is built to fail visibly rather than continue silently. Missing data, failed processes, and unmet conditions are surfaced as explicit errors — never suppressed or worked around.
EU data residency
The application and its data are hosted within the European Union. Technical documentation does not leave EU infrastructure in the course of normal operation, keeping processing within a single, known jurisdiction.
GDPR-aligned handling
Data is collected only where it serves a purpose, retained only as long as needed, and removed on request. Retention windows are explicit, and right-to-erasure is supported as a first-class operation, not a manual exception.
Authentication & access control
Accounts are protected with hashed credentials and email-based two-factor authentication. Access is role-aware, and no account can reach data it was not granted — confirmed at the application layer on every request.
Strict tenant isolation
Each organization's data is isolated from every other. Documents, audits, and findings are scoped to their owner, so one account's material is never visible or retrievable by another.
Encrypted at rest and in transit
Data is encrypted in transit over TLS and at rest on disk. Stored documents and database contents remain protected even at the storage layer, beyond application-level access controls.
Audit logging
Security-relevant actions — authentication, access, and changes to data — are logged. The record supports traceability and accountability, and underpins the usage and retention controls applied to each account.
Additional mechanisms include a 4-layer N/A determination model, cross-document consistency checks, and question-level re-run capability. The underlying model is constrained to retrieved document content — it cannot reference external knowledge or generate claims beyond what the evidence base contains. Every output is verified against source location, and remains draft until the reviewer confirms.
From upload to deletion
Every document uploaded to the Simulation module follows the same controlled path
with retention and erasure built in, not bolted on.
"A system you can trust with regulatory work has to be able to show its reasoning — and the sensitive data behind it deserves to be handled in one known jurisdiction, under clear controls, with deletion always on the table."
Structure what can be structured. Automate what can be automated. Keep the human in control.
Request an assessment
Questions about how the system is built or how your data is handled? Leave your details and we'll reach out.