Privacy · Last updated 2026-05-06

Your audio is yours.

We do not train models on your audio. Files are stored on Cloudflare R2 with a 24h lifecycle TTL and deleted automatically. We log only metadata (timestamps, durations, sizes, error codes) — never transcript content.

Sub-processors

• Vercel — hosting
• Cloudflare R2 — storage
• Supabase — database
• Modal — GPU compute
• Clerk — auth
• Stripe — payments
• Sentry — error monitoring
• HuggingFace — pyannote diarization model only (no audio sent)

Your rights

Email privacy@whisperly.dev for export or deletion.