Harmonai
Open-source generative audio tools by Stability AI
Founded
2022
Starting Price
Free
Category
AI Voice & Audio
Last updated March 8, 2026
About Harmonai
Harmonai is a Stability AI lab releasing open-source generative audio tools to make music production more accessible. Its flagship model, Dance Diffusion, uses diffusion technology to generate novel audio samples, enable style transfer, and interpolate between sounds. All outputs under the base models are released under CC0 license for unrestricted use.
Pros & Cons
Pros
- Completely free and open source with no subscription fees or hidden costs
- CC0 licensed outputs allow unrestricted commercial use of generated audio
- Backed by Stability AI infrastructure and research capabilities
- Active community of musicians, researchers, and developers
- Multiple generation modes: unconditional, style transfer, and interpolation
Key Features
Dance Diffusion
Generative audio model that creates novel sound samples using diffusion technology, producing 1-3 second audio clips
Audio Style Transfer
Transform existing audio recordings by applying the style and characteristics of trained models
Audio Interpolation
Blend and morph between two audio files to create unique transitional sounds
Text-to-Audio Generation
Describe the type of sound you want using text prompts with CLIP-like text encoding
Custom Model Training
Train your own models on specific audio datasets for personalized sound generation
Google Colab Integration
Run models directly in Google Colab notebooks without local GPU requirements
API Access
Developer API for building custom applications and integrations on top of Harmonai models
Pricing
Open Source
Free
- All generative audio models
- Dance Diffusion access
- Google Colab notebooks
- Custom model training
- CC0 licensed outputs
Best For
Sound Design for Music Production
Generate unique, never-before-heard samples and textures for electronic music, film scores, and sound libraries
Audio Prototyping and Experimentation
Quickly explore sonic possibilities by interpolating between sounds or applying style transfer to existing recordings
Sample Library Creation
Build custom royalty-free sample packs using CC0 licensed outputs for commercial distribution or personal use
AI Music Research
Study and experiment with generative audio models in an open-source environment for academic or commercial R&D
