Version 1.0 · in force
Subprocessors
Last updated: 19 September 2026
In force from 19 September 2026. We make no certification claims (for example ISO 27001 or SOC 2). Organisations using Haelyx for their staff can sign our data processing agreement.
These companies process data on our behalf to run Haelyx. We will update this page before adding a new subprocessor.
Whop
Purpose: sign-in, checkout, billing, memberships and DocSpeaks licence keys. Data: account identifiers, email, payment and membership records (card details are handled by Whop and its payment processors, never by Haelyx). Location: United States. Safeguards: the provider's standard data processing agreement; transfers under Standard Contractual Clauses or the EU–US Data Privacy Framework where required.
Vercel
Purpose: website hosting and server functions. Data: requests to the app, including text sent for reflection processing, in transit. Location: functions configured for Frankfurt, Germany (fra1); company based in the United States. Safeguards: the provider's standard data processing agreement; transfers under Standard Contractual Clauses or the EU–US Data Privacy Framework where required.
Neon
Purpose: database. Data: account records, encrypted reflections, sessions (hashed), rate-limit counters and audit log. Location: AWS eu-central-1, Frankfurt, Germany. Safeguards: the provider's standard data processing agreement; data stays in the EU.
Language model provider
Purpose: generating reflection questions and patterns in the signed-in app, and rewriting debrief questions in Haelyx Debrief when you press “Send to the model”. Data: case description, questions and answers for that request; for Haelyx Debrief, the scenario title, objectives, roles and tagged notes of a simulated scenario. Provider and location: Mistral AI (Paris, France), reached through Vercel AI Gateway. Used by the signed-in Case Reflection Engine and, only on that button press, by Haelyx Debrief. EU-based provider; our requirement is no training on your text and minimal retention. DocSpeaks Web does not use this provider: its speech model runs on your own device.