Riffusion Pricing & Plans

Riffusion uses a Stable Diffusion model trained on spectrograms, which are visual representations of audio showing frequency and amplitude over time. When you enter a text prompt, the model generates a corresponding spectrogram image that is then converted back into playable audio using an inverse short-time Fourier transform (ISTFT). This unique visual-to-audio approach allows it to transition smoothly between different musical styles.