Meetings are among the most sensitive things a company records. Peaflow treats them that way — isolated per workspace, access-controlled at the database, and never fed into model training.
Every row is scoped to your workspace and enforced by database row-level security — one customer can never read or write another's meetings.
OAuth tokens and recordings are held server-side only, behind deny-by-default policies. Recordings are served through short-lived signed links.
Every inbound event — meeting, billing — is signature-verified and idempotent, so nothing forged or replayed is ever processed.
The notetaker appears in the participant list and the host admits it — everyone in the call can see it's recording.
Your transcripts and recordings are used to serve you — summaries, search, recaps — and never to train foundation models.
Set how long recordings are kept, delete any meeting on demand, and account deletion anonymises your data rather than orphaning it.
In managed Postgres (Supabase) with row-level security, and recordings in Cloudflare R2 accessed through short-lived signed URLs. Peaflow never relies on a meeting provider's storage as the permanent copy.
No. Isolation is enforced at the database with a membership check on every read and write. We regularly run second-tenant attack tests against production to prove it.
No. Your content is used only to generate your own summaries, answers and recaps.
Payments run through Polar with signature-verified, idempotent webhooks. Card details never touch Peaflow's servers.