Privacy Policy

Last updated: May 9, 2026

What we collect

Flow processes your speech to convert it to text. To do this we collect:

• Your audio while you actively dictate (key held / hands-free toggled on)
• The transcribed text we return to you
• Your account email and authentication identifier
• Aggregated usage metrics (words dictated, audio duration) to enforce tier limits

What we don't collect

• Audio when you're not actively dictating — Flow only listens while the hotkey is held or hands-free is toggled on
• Anything you type without using Flow
• The contents of other apps on your computer
• Payment card details — those are handled exclusively by Stripe

How we process your audio

Audio is sent to Groq Inc. for transcription via Whisper Large v3 Turbo, then to Groq again for cleanup using Llama 3.3 70B. Both pass through our Cloudflare Workers proxy. We do not store your audio after transcription completes — the audio file is discarded once the cleaned text is returned.

Transcribed text history is stored locally on your device by default. Pro users may opt into cloud sync of snippets and dictionary entries (encrypted at rest in Supabase Postgres).

Subprocessors

We rely on these vendors to operate Flow:

• Groq Inc. — speech-to-text + cleanup
• OpenAI — fallback speech-to-text + cleanup
• Cloudflare — proxy infrastructure
• Supabase — authentication + database
• Stripe — payments

Your rights

You can request export or deletion of your data at any time by emailing privacy@flow.app. We respond within 30 days.

Contact

Questions? hello@flow.app