Purchases and access
Restore purchases in the Premium screen. Apple manages billing, renewal, cancellation, and refunds. Deleting your Stable Audio data does not cancel a subscription.
Stable Audio support
Review the common topics below. If you still need help, compose a private email from your own mail app and include only the information needed to investigate the issue.
Restore purchases in the Premium screen. Apple manages billing, renewal, cancellation, and refunds. Deleting your Stable Audio data does not cancel a subscription.
There is no daily limit. Plan allowances can be used at any time in their billing period, while separately purchased credit packs do not expire.
Include the selected model, workflow, duration, output format, and Generation Record ID. Do not send API keys, payment details, or unnecessary source audio.
Use an available Lyria model for structured vocals and custom lyrics. Stability models may create vocal textures but do not guarantee exact words.
Record or import a vocal for a local studio mix, or explicitly choose AI analysis to transcribe the words and detect your musical delivery, tempo, and style. Review the editable draft, then choose an available vocal model. The AI song uses a new vocalist and does not preserve or clone your recorded voice. The analysis cost is shown before you start; the reserved credits are captured for a usable analysis and released if analysis fails.
The analysis copy is a temporary mono WAV limited to five minutes and 30 MB. It is deleted from upload storage and Google Files after analysis. Local mixing remains on your device when you do not choose AI analysis.
With Lyria 3.5, you may add selected images to guide your music. Only the images you explicitly choose are sent through our secure backend to Google. Photo references are optional, and you must have permission to use them.
Upload only MP3 or WAV audio you own or are authorised to transform. Streaming pages, DRM-protected media, and YouTube, Spotify, or Apple Music links are not supported.
Improve with AI fixes only necessary grammar, sentence, logic, or consistency issues and translates supported prompts to English without adding creative details. This text-only processing uses OpenAI; audio is never sent to OpenAI.
Ready results can be replayed and downloaded in their original MP3 or WAV format. The Generation Provenance Record is factual documentation, not a licence certificate.
Completion alerts, twice-weekly inspiration reminders, and Analytics & Diagnostics are optional controls. They can be changed in Stable Audio Settings or iOS Settings.
Version 1.3 and later have no ads, advertising identifiers, or cross-app tracking. The rewarded-ad trial has ended and new sessions are disabled for earlier versions too. Music Credits already earned remain available. Update the app to remove its old advertising components.
Contact the developer
Your controls
Stable Audio Settings lets you export the service record, clear settled history, withdraw AI-processing permission, or permanently delete the pseudonymous account and associated data. Manage any Apple subscription separately in App Store settings.