Service without touching patient data
An AI talking to patients cannot exist without safety. In Medvix, every message passes through consent, masking, and risk control first.
Medvix is not a doctor and does not provide medical diagnosis or treatment. In emergencies, call 112.
How it works
Consent-first
No chat starts without KVKK notice and explicit service consent from the patient.
PHI masking
National IDs, phone numbers, emails are masked as early as possible and never stored in plain text.
Fail-closed
When medical advice, drug recommendations, or identity collection risks are detected, the AI falls back to safe boundaries.
Emergency routing
When an emergency is detected, a 112/ER banner is shown. The AI never encourages continuing chat instead of seeking care.
Decision flow in three steps
Consent → AI decision → reply or human handoff. Every step is logged in the audit trail.
Try it yourself
See masking and decision behavior across scenarios. (Synthetic data; no real patient information.)
Compliance
- KVKK (Turkish Data Protection Law 6698) compliant
- Data Processing Agreement (DPA) ready to sign
- Patient data stays on servers in Turkey
- Audit log retained for every message
- Annual penetration testing and security assessment
Let's review together
We can align security rules for your clinic in 15 minutes.