How we protect your data, your clients' data, and your trust. Enterprise-grade security with full transparency.
We use the same encryption standards trusted by banks, hospitals, and government agencies.
All stored data -- OAuth tokens, client configurations, and email metadata -- is encrypted with AES-256-GCM, the gold standard in symmetric encryption.
Every connection between your browser, our servers, and third-party services is encrypted with TLS 1.3 -- the latest transport layer security protocol.
Your data is yours. Nobody else's system can see it, touch it, or query it.
Every database table enforces row-level security policies. Your data is partitioned at the database level -- not just the application level.
Each client's data is scoped by ownership checks on every request. No cross-client data leakage is possible, even if credentials are compromised.
Database access is restricted to service_role credentials. Anonymous and authenticated public access has been explicitly revoked.
We are custodians of your data, not owners. You have full control at all times.
All data generated through AssistantAI -- drafts, classifications, contact records, preferences -- belongs to you. We make no claims to it.
Request a complete export of your data at any time. We provide it in standard formats (JSON/CSV) within 24 hours.
Your data is never sold, shared, or used for advertising. It is used solely to power your AI assistant and improve your experience.
We use best-in-class infrastructure providers. Here is exactly what each one accesses.
| Service | Purpose | Data Accessed | Security |
|---|---|---|---|
| Supabase | Database & storage | Client records, configurations, email metadata | Security page |
| Gmail API | Email access | Email content (read/draft/send via OAuth) | Security page |
| Claude API | AI processing | Email content for classification & drafting | Security page |
| Stripe | Payment processing | Billing info, subscription status | Security page |
| Resend | Email notifications | Recipient email, notification content | Security page |
AI drafts responses in your voice, but every email waits for your review. You see the draft, edit if needed, and tap approve. If you don't approve it, it doesn't send. You are always in control.
We follow industry best practices and are building toward formal certifications.
Join Phoenix professionals who trust AssistantAI with their email.
Get Started