Your voice recordings and transcripts are processed by OpenAI's API and then forgotten. They never become training data for any model — yours, ours, or theirs.
Recordings are streamed to Whisper for transcription and immediately discarded. We keep the resulting text, not the audio.
Your account sees only your data. We use a strict user-scoped database with auth on every endpoint. Your insights cannot leak to another user.
Every session and reading has a delete button. Hit it and the data is gone — from the database, irreversibly. There is no "soft delete" backup.
Knack uses OpenAI's API for transcription and analysis. We trust their enterprise data policy: 30-day retention for abuse monitoring, no training, then deletion. We don't see, store, or share your recordings.
Install Knack as an app — faster open, full-screen, ready when you need a sharper ear.
Made with Emergent