Sub-processors

Clinical and patient-data processing

These tasks use the necessary request payload in plaintext while the request is active. The Maestra relay does not log or retain the clinical payload.

Service	Current purpose	Necessary payload	Safeguard recorded in the counsel-reviewed ROPA
Modal Labs	Audio transcription, speaker processing, text embeddings, and retrieval reranking	Audio chunks or text being embedded or ranked	Executed DPA with GDPR or UK SCCs relied on under the documented equivalence basis
Eden AI	Gateway for generated notes, assistants, transcript-processing tasks, and patient-import parsing	Clinical text or source text, images, and PDF documents required for the task	DPA in place for the documented EU inference path

The page intentionally does not name a downstream language model because runtime routing can change without changing the processor relationship.

PDF rendering

Generate PDF and the patient-record export render each patient PDF by sending clinical HTML transiently to the Maestra relay renderer. JSON serialization and final ZIP assembly happen locally. The HTML is not written to the stored database by the renderer.

Account and platform processing

Service	Purpose	Data processed
RevenueCat	Subscription entitlement management	Account identifiers and purchase records
Stripe	Web subscription payment and management	Payment details and account email
Apple and Google	Native authentication and store subscriptions	Platform account and purchase data
Resend	Authentication and account emails	Email address and message delivery data
Sentry	Error monitoring	Anonymized crash and error data without plaintext clinical content

For general questions about a processor, contact [email protected]. Statutory privacy requests use the DPO contact listed in the applicable privacy framework pages.