GDPR is not a checkbox for EU therapists — it is the legal framework governing every piece of patient data you handle. This page explains how Freudche is designed to work within that framework.
GDPR-first · EU data architecture · Observer-level AI
Many AI tools for therapists are built for the US healthcare market and adapted for Europe. HIPAA and GDPR are different legal frameworks for different jurisdictions — a HIPAA-compliant tool is not automatically GDPR-compliant. EU therapists need tools built within their regulatory context from the start.
Applies to all data subjects in the EU/EEA · Broad scope including therapy session data · Right to erasure, portability, restriction · High standards for health data (special category) · DPA enforcement per EU member state
Applies primarily to US covered entities · Different consent and breach notification standards · International transfers require separate GDPR compliance · Does not replace GDPR for EU data subjects
This is a general overview for context — not legal advice. Consult a qualified legal professional for your specific situation.
Freudche uses Gemini (Google Cloud) for AI analysis. Session transcripts are anonymised before leaving your network — speakers are passed as [Speaker 0] and [Speaker 1], never by name.
Transcript text never leaves your network for vector search. Freudche computes embeddings locally using sentence-transformers, eliminating the need to send raw session content to external embedding APIs.
Speakers are passed to external AI services as [Speaker 0] and [Speaker 1] — never by name. This significantly reduces the personal data that external AI services process.
Where data transfers to processors outside the EU are necessary (transcription service), Freudche uses Standard Contractual Clauses as a transfer mechanism. Formal Data Processing Agreements with sub-processors are in progress.
Every AI observation passes through two validation layers before reaching the database: a deny-list scanner blocking diagnostic labels, risk screening terms, and pathologising language — followed by a secondary LLM audit. Only observer-level observations reach the therapist.
Every AI-generated note carries a machine-readable marker confirming it was AI-generated, in line with EU AI Act Article 50 transparency requirements. Freudche is internally classified as a high-risk AI system — enforcement is expected from December 2027.
Freudche never uses session data to train AI models. Session data is used solely for direct clinical documentation support.
Freudche generates observations, not diagnoses. A two-layer compliance engine actively blocks diagnostic labels, risk screening claims, and pathologising language from reaching the therapist. The therapist retains full clinical responsibility.
Under Dutch law (WGBO) and the professional codes of BIG-registered therapists, the duty of professional secrecy (beroepsgeheim) rests with the therapist — not with the tools they use. Freudche operates as a processing tool under the therapist's responsibility. Using an AI tool does not transfer or reduce the beroepsgeheim obligation. Therapists remain responsible for: obtaining informed consent from patients before using AI-assisted documentation; the careful handling and clinical evaluation of all AI-generated outputs; and ensuring any data processing agreement with Freudche is in place before using the service with real patient data.
Sharper sessions. Faster clarity. Your patients feel the difference — you see the evidence. Book a live demo and see it on real session work.
Book a live demoWe use privacy-friendly analytics cookies to understand how the site is used and make it better. Choose "Accept" to allow them, or keep only the cookies needed to run the site. Privacy Policy