RegentDesk← Compliance overview

Work in progress

This page is a work in progress and subject to changes and rewrites. Content here reflects RegentDesk's current intent — language, commitments, and structure may all change before launch.

Subprocessors

Every third-party vendor that processes advisor or client data on RegentDesk's behalf is listed here. RegentDesk commits to giving advisors at least 30 days' notice before adding any new subprocessor that processes client data.

Last updated: April 24, 2026

Active subprocessors

Subprocessor
Purpose
Data categories processed
Location
Supabase
Database, file storage, and auth storage
All client data, drafts, voice samples, life events, interaction history, advisor profile
United States
Clerk
User authentication and session management
Advisor email address, identity tokens, session metadata, MFA state
United States
Anthropic
AI draft generation and context extraction
Email content and client context submitted during generation. Not retained beyond the API call. Not used to train models.
United States
Vercel
Application hosting and serverless function execution
All data flowing through the application in transit. Logs retained per Vercel's standard policy.
United States
Inngest
Background job queue for batch email generation
Job metadata and payload (batch job IDs, client IDs, status). Does not store email content.
United States
Stripe
Payment processing and subscription management
Billing data only: payment method metadata, subscription state, invoice history. No client data.
United States

Planned subprocessors

These vendors are listed in advance of integration. They are not active today. Advisors will be notified at least 30 days before activation.

Subprocessor
Purpose
Data categories
Location
Google Cloud Pub/Sub
Gmail push notifications (future Gmail integration)
Gmail message IDs and change notification metadata. No email body content.
United States
Microsoft Graph
Outlook push notifications (future Outlook integration)
Outlook message IDs and change notification metadata. No email body content.
United States

A note on Anthropic

Per Anthropic's API terms, inputs sent to the Anthropic API are not retained beyond the duration of the API call and are not used to train Anthropic's models. When RegentDesk generates a draft, the content is processed in real time and the result is returned. Anthropic does not store a copy. This commitment is documented in our Data Processing Agreement template, available on request.

How we notify you of changes

Before adding any new subprocessor that processes client data, RegentDesk will email all active advisors at least 30 days in advance. The notification will include the subprocessor name, purpose, data categories, and geographic location. Advisors may object to the addition; the process for objection and the available remedies are documented in the Data Processing Agreement. This page will be updated to reflect all changes.