Stable Audio 2

What Stable Audio is

Stable Audio is Stability AI's generative audio platform. It produces music and sound effects from text prompts or from audio you supply, and version 2 extended output to full tracks of up to three minutes with clear musical structure and stereo output at standard audio resolution. Audio-to-audio conversion and style reshaping are included, and an open weights release exists for local experimentation.

What you can do with it

• Generate music and effects from prompts
• Convert one piece of audio into another style
• Produce longer structured tracks
• Run the open model locally where licensing allows
• Prototype sound design ideas

Who it is for

• Music and film producers
• Game sound designers
• Researchers and local model users

What to watch out for

• Licence terms differ between the hosted service and the open weights release; check both before commercial use
• Training data provenance for generative audio is contested across the industry, so follow current legal developments
• Longer tracks still need mastering for release
• Local execution needs capable hardware