Features
- Open-source Dance Diffusion models generate short audio clips, drum loops and melodic textures
- Trained on legally cleared datasets
- Generates custom sound libraries of kicks, snares and synth stabs
- Sample-to-sample transformation that re-imagines a beat or melody in another genre or instrument
- Infinite non-repetitive loops for games and streams
- Batch generation of sample packs ready for DAWs
- Latent space exploration between instruments
- Research powering the Stable Audio platform
- Code and models available on GitHub and Hugging Face
Pricing Details
💵 Pricing (2026 Status) Harmonai is primarily an open-source organization, but its technology is monetized through Stability AI’s commercial arms. Access Type Price Key Features Open Source (GitHub) Free Download models (like Dance Diffusion) and run them locally for free. Hugging Face Free / Credit-based Web-based testing and generation using public community spaces. Stable Audio (Pro) $9.99 - $29.99 /mo The "commercial" version with a polished UI, longer generation times, and commercial licenses. 🎁 Is the free version a trial or completely free? The core Harmonai models are completely free and open-sourcefor research and personal use. If you have the technical skill to run them on your own computer (GPU required), you never have to pay. For commercial "ready-to-use" features, users typically migrate to Stable Audio.