Privacy for artists uploading unreleased music

Two categories of processing are involved. First, the managed application backend: it handles your account, the database records describing your work, and the private storage holding your audio. Second, specialist compute: analysis, stem separation, dataset preparation, training and generation are computationally heavy and run on connected services configured for your studio, which receive the audio required to perform the requested job.

This is disclosed rather than hidden because it matters to your assessment. If that connected service is not configured, ANYANO tells you the capability is unavailable — it does not reroute your audio elsewhere or fabricate a result.

Data transfers are encrypted in transit. Retention and jurisdiction for each processor are governed by the Privacy Policy and the relevant providers' terms.

We will not describe the service as "100% private" or "fully offline". Your audio is processed by systems we operate and by connected compute services; that is how training and generation happen at all. What we can say precisely is that it is not published, not shared with other accounts, and not used to train other customers' Sounds.

We also will not claim that deletion is instantaneous everywhere. Operational backups exist for disaster recovery, and deleted material can persist in snapshots until they roll off on their normal schedule.

If a specific recording is commercially sensitive and you are still evaluating the product, start with material you are comfortable processing — older demos, stems from released work, instrumental beds — and expand once you have seen how the pipeline behaves. A Sound trained on representative material tells you most of what you need to know.

Keep your own masters. ANYANO is a creative tool, not an archive: maintain your primary backups independently of any service.