Viznerve

Trust & security

Your accounts, protected

We take the security of your connected accounts seriously and design so that a breach yields unreadable data.

Built so a breach yields unreadable data

We don't claim to be “breach-proof” — no one honestly can. Here's what we do instead.

  • We never store your Gmail loginInbox access is brokered by Composio — your Google password never touches our servers.
  • Everything else is encryptedConnected-account tokens are encrypted with a per-tenant key. If our database were stolen, they stay unreadable.
  • Connect or disconnect anytimeDisconnecting an account erases the data we stored for it. You stay in control.
  • Your studio's data is isolatedRow-level isolation keeps each studio's data separate, and your content never trains any AI model.

How it works

Encryption in transit & at rest
TLS on the wire; connected-account tokens and sensitive fields encrypted at rest with a per-tenant key.
Tenant isolation
Row-level security keeps each studio's records separate at the database layer — not just in the app.
AI data handling
Inquiry text is sent to Anthropic (Claude) to classify and draft; your OAuth tokens never enter any AI context, and your data is not used to train models.
Access & recovery
Sign in by magic link with optional two-factor (TOTP); encrypted off-site backups; a credential kill-switch and an access audit log.
You're in control
Connect or disconnect any integration anytime — disconnecting erases the data stored for it. Export or delete your studio's data on request.
A shared responsibility
We secure the platform — isolation, encryption, backups. You secure your account — your sign-in and the team access you grant. Turn on two-factor, and tell us fast if something looks off.

We are not currently SOC 2 or ISO 27001 certified, and we will pursue certification as we grow. We'd rather tell you that plainly than imply a guarantee we can't keep.

Privacy PolicyTerms of ServiceReport a vulnerability

Responsible disclosure: email [email protected]. We ask for reasonable remediation time and won't pursue good-faith research.