Stable Audio Open
Stable Audio Open 1.0 is Stability AI's text-to-audio model for local use through stable-audio-tools or Diffusers, generating up to 47 seconds of 44.1 kHz stereo; its model card says it performs better on sound effects and field recordings than on complete music.
Why use it
It is selected because teams can pin the model, prompt, seed, and post-processing version, batch footsteps, machinery loops, ambience layers, and abstract textures offline, and connect inference to their own review and asset-naming pipeline; this is stronger than a browser quota when experiments must be reproducible, confidential material must stay local, or many candidates are required.
Where it fits
Use art, animation, audio, levels, and narrative to make the game easy to read and its world convincing.
What to check
The weights use the Stability AI Community License rather than a standard open-source license: commercial users must register, annual revenue above USD 1 million requires a separate enterprise license, and distributing the model or derivatives brings Notice, agreement-copy, and Powered by Stability AI attribution duties. MIT tool code does not override weights, AUP, training-data, or output risks, and local use still needs adequate VRAM plus loudness and loop review.