Politique de confidentialité

⚠️ BROUILLON — À FAIRE VALIDER PAR UN AVOCAT AVANT PUBLICATION. The text below is a draft. Get a real legal review before publishing publicly.

Short version:Your audio, transcripts, and summaries belong to you. We don't train AI models on them. We don't sell your data. You can delete everything any time.

1. Who we are

Transkribe is operated by [Legal Entity Name], registered in [Jurisdiction]. This policy explains what data we collect, how we use it, and your rights.

2. What we collect

• Account info: name, email, avatar — provided by Google when you sign in.
• Files: audio you upload, video you upload (we discard the video and keep only audio), in-browser recordings.
• Transcripts and summaries: generated from your files.
• Usage logs: what features you use, when, how often.
• Billing info: handled by Solid — we never see card numbers.

3. What we never do

• Train AI models on your audio, transcripts, or summaries.
• Sell your data to third parties.
• Read your transcripts (humans on our team don't have access unless you give us permission for support).

4. Vendors we use

Your files and metadata are processed by:

• Cloudflare R2 — storage of audio files and exports.
• Replicate — running Whisper for transcription and pyannote for diarization.
• Google (Gemini) — generating AI summaries.
• Solid — payments.
• Clerk — authentication.
• Sentry — error tracking (no PII).
• GTM / GA4 / Amplitude — usage analytics (after consent).

5. Your rights

You can:

• Access your data — request an export from your profile.
• Delete your data — delete files individually or your whole account (30-day grace).
• Object to certain processing — contact us.

6. Contact

Questions about your data? Email privacy@transkribe.com.