
AI music generation from text
Riffusion is a free, open-source AI music generation tool that creates music from text prompts using a novel approach based on Stable Diffusion applied to audio spectrograms. It generates short music clips in real-time, allowing users to explore a vast range of musical styles and ideas through simple text descriptions.
Riffusion generates music by converting text prompts into spectrogram images using a fine-tuned Stable Diffusion model, then converting those spectrograms into audio. This unique approach allows it to create music spanning virtually any genre or style describable in text. Users can interpolate between two different prompts to create smooth musical transitions, generate continuous music streams, and explore creative blends of styles. The tool runs in a web browser and produces results in seconds. As an open-source project, the underlying model and code are freely available for developers and researchers.
Riffusion is perfect for music enthusiasts, creative experimenters, and developers interested in AI music technology. Its free, open-source nature makes it ideal for researchers studying AI music generation, hobbyists who want to explore text-to-music concepts, and developers looking to build on top of AI music technology. It is more of an experimental tool than a production music solution.
Visit riffusion.com to start generating music immediately in your browser with no account required. Type a text prompt describing the music you want (e.g., "jazz piano in a rainy cafe" or "epic orchestral battle music"), and the AI generates a clip in seconds. Experiment with the interpolation feature to blend two prompts together. Developers can access the open-source code on GitHub.
Pricing & Accessibility: Riffusion is completely free to use on the web. The open-source model and code are available on GitHub under a permissive license. No account or subscription is required. The web app runs in any modern browser.
Why Consider Riffusion: Riffusion offers a completely free, no-account-needed AI music generation experience with its innovative spectrogram-based approach, making it the most accessible entry point for exploring text-to-music AI technology.
Exploring AI music generation concepts, creating short musical sketches and ideas, experimenting with genre-blending through prompt interpolation, researching AI audio technology, generating quick musical prototypes for creative inspiration
Free
Free tier: Fully free with no limitations