Riffusion
by Riffusion
Stable Diffusion for audio, and the cheapest way to sketch a loop.
Riffusion turns a prompt into a few seconds of instrumental audio in about the time it takes to read this. Short by design: it is for finding a direction, not for finishing a track.
More details
Riffusion by Riffusion is a fast-speed AI music generation model. Riffusion turns a prompt into a few seconds of instrumental audio in about the time it takes to read this. Short by design: it is for finding a direction, not for finishing a track. It renders instrumental tracks, clips of 8 seconds, 16 seconds, prompts up to 600 characters.
Riffusion is priced at $0.0005 per second of generated audio, so the default 8-second clip costs $0.004 and the longest 16-second clip costs $0.008. Commercial use of the generated audio is allowed under the provider terms. You can generate music with Riffusion on idapt.app alongside 200+ AI models in one workspace.
Pricing
Cost per second
$0.0005
Default clip (8s)
$0.004
Billed on the length of the rendered audio
Longest clip (16s): $0.008
Capabilities
Speed
Fast: Short effects in seconds
Output
instrumental
Lyrics
Instrumental only
Lengths
8s · 16s
Max Prompt
600 characters
Commercial use
Allowed by the model provider
Riffusion questions
Who makes Riffusion?
Riffusion is made by Riffusion and is served on idapt through the Riffusion API.
How much does Riffusion cost?
$0.0005 per second of generated audio. A 8-second clip costs about $0.004.
Does Riffusion produce vocals?
No. It renders instrumental audio only, with no sung or spoken voice.
How long can a Riffusion clip be?
Lengths are picked from a fixed set: 8 seconds, 16 seconds. The default is 8 seconds and the longest is 16 seconds.
Can I use Riffusion output commercially?
Yes. The provider terms allow commercial use of the generated audio.
What are the alternatives to Riffusion?
idapt lists 12 music generation models. The most-used alternatives are MusicGen Stereo Large and MiniMax Music 2.6.
A 8-second clip on this model costs about $0.004.