AethonVoice processes your data to deliver audio and nothing more. This page explains exactly what happens to your data, how long we keep it, and how we protect it.
AethonVoice does not use your input text, generated audio, or voice clone references to train or improve any AI model.
Your data is processed, delivered, and deleted. It is never fed into a training pipeline, never used to improve voice quality for other users, never shared with third parties. The underlying OmniVoice model is open-source and was trained exclusively on public datasets before AethonVoice existed.
| Data Type | Retention | Details |
|---|---|---|
| Input text | Deleted immediately | Text is held in memory only during audio generation. Once the job completes or fails, the input text is purged. It is not written to persistent storage. |
| Generated audio | 7 days | MP3 files are stored in Google Cloud Storage with signed download URLs. Both the URL and the file are automatically deleted after 7 days. |
| Voice clone references | Persistent (user-managed) | Reference audio for custom voice cloning is stored in your account's voice library. You can view, use, and delete cloned voices at any time. |
| Job metadata | 30 days | Job ID, status, timestamps, language codes, voice name, and audio duration are retained for 30 days for debugging and usage tracking, then automatically purged. Job metadata does not include input text content. |
AethonVoice logs the minimum metadata necessary for service operation.
| Layer | Implementation |
|---|---|
| In transit | All API traffic encrypted via TLS 1.2+ (HTTPS only) |
| At rest | Generated audio and voice references encrypted at rest (AES-256) |
| Authentication | API key required for all requests. Keys are hashed before storage — AethonVoice cannot see your raw key after creation. |
| Download URLs | Signed URLs with 7-day expiry. No authentication required to download — the cryptographic signature carries access. URLs cannot be guessed or enumerated. |
| Voice clone data | Isolated per account. One account cannot access another account's voice library. |
Babbly AI Company Limited
559/67 Thanapat Haus, Nonsi Road, Yannawa, Bangkok, 10120, Thailand
Contact: aethon@lab.ai
All data processing and storage occurs within US-based infrastructure.
AethonVoice complies with Thailand's Personal Data Protection Act. Same rights as GDPR: access, correction, deletion, and portability.
For privacy-related inquiries: aethon@lab.ai
Delete individual cloned voices via the API or dashboard at any time.
Request full account deletion via support — removes all job history, voice library, and account metadata.
Generated audio files are automatically deleted after 7 days regardless of account status.
Your data stays yours. Start building with AethonVoice.
