Who Can Access Your Data

The short answer: your canvases are private to your account. There is no sharing, no team access, and no public link mechanism in the Beta — nothing you build can be seen by another user.

The technical answer:

ISOLATION. Every canvas row is stamped with your user ID at creation and every read is filtered by it. Requesting another user’s canvas URL returns “not found” — the server never loads it. There is no cross-user query path in the application.

AUTHENTICATION. Sign-in is handled by Clerk (clerk.com), a dedicated authentication provider. PIVVA never sees or stores your password.

STORAGE. Data lives in a managed PostgreSQL database (Neon, hosted on AWS US East / N. Virginia), encrypted in transit (TLS) and at rest (AES-256). The application is hosted on Vercel; traffic to pivva.io is HTTPS end to end.

WHAT WE DON'T DO. Your models are not used to train AI. There is no analytics or tracking script in the app. No service receives your canvas data automatically; the only third-party transmission that exists is one YOU trigger — the Send Feedback button, which sends your message (and a screenshot only if you choose to attach one) to our feedback system. The infrastructure named above processes your data solely to run the service.

OPERATOR ACCESS. PIVVA is run by its founder, who — like any database operator — has administrative access to production data for operating and supporting the service. That access is used for support and debugging, with your knowledge where your specific data is involved, and for nothing else. Founding Members putting client data into PIVVA can request a short written confidentiality note — just reply to your invitation email.

If your work requires more (a DPA, data deletion, an export of everything you’ve stored), ask. At this stage you are talking directly to the person who can do it.