Legal

Policies

AI & Voice Data Policy

Last updated June 2026 · Powered by SoundForgeArena

Our engine

Text understanding, Production DNA and composition planning are powered by AuraSession's proprietary engine. Voice briefs are converted to text by a confidential speech-to-text sub-processor. We do not publish the underlying model details.

What is sent

Only the brief text (and, for voice, the audio clip) needed to perform the operation you requested is sent to the model provider. We never send your payment details to AI providers.

Retention

Voice audio is discarded after transcription. Generated artifacts are stored so you can download and push them to your DAW; you can delete projects at any time.

No hidden training

We do not use your private creative content to train our own models.

Honesty

Aura Vision events reflect real operations. We never fake generations, progress, DAW connection, revenue, or completion.

Desktop-first studio

AuraSession works on your phone, but full sessions — MIDI, DAW packages and sample matching — are built for Mac & desktop. Open on a computer for the real thing.