Stable Audio Open
Open text-to-audio diffusion model for generating short instrumentals, loops and sound effects locally.
| Language | Python |
|---|---|
| Category | ML & generative |
| License | MIT |
| Platforms | macOS Windows Linux |
| Install | pip install diffusers transformers torch soundfile |
| First released | 2024 |
| Maintained | Yes |
| Links | Website · Source |
Strengths
- Open weights you can run on your own GPU
- Strong at drum loops, textures and sound design
- Also usable via the diffusers StableAudioPipeline
Limitations
- Instrumental only — no vocals
- Model weights use the Stability AI Community License, which restricts large-scale commercial use
- Generates short clips (tens of seconds)
Best for
Generating instrumental loops and sound effects locally with open weights.
Minimal example
from diffusers import StableAudioPipeline
import torch, soundfile as sf
pipe = StableAudioPipeline.from_pretrained('stabilityai/stable-audio-open-1.0', torch_dtype=torch.float16).to('cuda')
audio = pipe('128 BPM tech house drum loop', num_inference_steps=200, audio_end_in_s=10.0).audios[0]