FUZZ-1.1 is a legacy music generation model in the Riffusion FUZZ series, suitable for turning song themes, style ideas, or self-written lyrics into listenable audio. It offers both inspiration and custom creation modes, and can also be configured for instrumental tasks. On this platform, it can be accessed through the Riffusion or Producer audio entry points, with works and subsequent tasks managed around song IDs.
Submit requests to the public API at api.acedata.cloud using the documented parameters, then use the results in your application.
Input parameters and result formats vary by service. Use the public API for this model and follow its guide for generation, task retrieval and editing operations.
Clarify capacity, inputs and outputs, and invocation methods before choosing a model.
The above are the input and output specifications for this platform's creation entry points. Use editing actions and control parameters according to the selected entry point and the applicable scope of the model.
Learn what FUZZ-1.1 can bring to your work.
When you do not yet have lyrics, you can first use prompt to describe the song's theme, mood, instruments, and stylistic direction, then initiate creation with generate. For example, organize “rainy night, soft piano, warm atmosphere” into a prompt to obtain listenable musical material and assess whether the idea is worth developing further.
When you already have lyrics, you can use custom=true to enter custom mode, put the lyrics in lyric, and add a title. In the Riffusion entry point, you can also fill in style separately, allowing lyric content and sonic direction to be expressed independently. This is suitable for first defining the narrative and chorus, then trying song drafts in different styles.
FUZZ-1.1 creation results are delivered as audio links and song information, making them easy to preview, save, and associate with projects. Applications can record song IDs, titles, and statuses, and use asynchronous tasks or callbacks to arrange result delivery, integrating music creation into content management workflows instead of always waiting for a single request to finish.
Start with specific tasks to find where the model can be effective.
Enter completed verses, choruses, and a song title, then add the styles you want to try to generate listening drafts for discussion. You can use the same lyrics to try different sonic directions, compare how the words are expressed in music, then decide which sections need adjustment; the deliverables are audio and the corresponding song record.
For short films, product demos, or event videos, enter descriptions of mood, instruments, and atmosphere, and use instrumental to set up an instrumental music task. First use generated audio to determine the soundtrack direction, then adjust the length and transitions in editing software; this approach is suitable for finding creative ideas, but prompts should not be treated as precise timelines.
For music projects that previously used FUZZ-1.1, you can continue explicitly selecting this model while retaining lyrics, prompts, and titles as creative records. Listen to new audio alongside older works and compare whether the expressive direction suits the current project; version comparisons should be based on the actual sound rather than the model name alone.
Choose based on task complexity, input materials, and expected results.
FUZZ-1.1 and FUZZ-1.1 Pro are separate models, and they also differ from FUZZ-1.0 and FUZZ-2.0. If you already have FUZZ-1.1 creation records, or want to continue experimenting with this historical version, you can choose it; for new projects, use the same lyrics and prompts for comparative listening, choose based on song expression, and do not infer results from version numbers alone.
If you want to fill in style separately, choose /riffusion/audios; if you already have a Producer integration workflow, use /producer/audios and explicitly set FUZZ-1.1. Requests for the two cannot be exchanged as-is; Producer requires action, lyric, and prompt together. If you need variation or editing tasks, you should also distinguish between entry-point actions and model applicability.
Arrange the inputs first, then connect them to the appropriate application workflow.
Prepare one set of original lyrics and two sonic directions, using a consistent title and main creative parameters.
Submit your creation using the exact FUZZ-1.1 call value; select the prompt, lyric, and title fields for the corresponding entry point, and save the task_id and work ID. Complete listening first, then choose follow-up actions according to this model's documentation, without mixing editing fields from other versions.
Design tasks directly from the following inputs and acceptance criteria.
Use the same lyric to compare a lighter acoustic option with a stronger drum arrangement; explicitly specify FUZZ-1.1, and save the result of each task.
Check the differences in verses, choruses, and lyric expression, and select the version suitable for the project; do not treat the version number as the sole basis for ranking audio quality.
Before formal use, understand the output quality and capability scope.
Answers to common questions when using FUZZ-1.1.
No, they are separately listed models. Calling FUZZ-1.1 does not make it equivalent to Pro because of their similar names. When choosing, you can compare them by previewing the same lyrics and style, then decide based on the results; do not apply Pro descriptions or generation examples from other versions to FUZZ-1.1.
Yes, you can start with inspiration mode and use a prompt to express the theme, mood, and sound direction, without completing the full lyrics first. However, requests must still include the fields required by the endpoint: Riffusion requires action and lyric, while Producer also requires prompt; whether lyric content is needed and whether a field is required are two different things.
Write the lyrics in lyric, set custom=true, and use action=generate and model=FUZZ-1.1. You can add title; in the Riffusion endpoint, you can also fill in style. It is recommended to organize lyrics into sections such as verses and choruses, then focus on previewing the lyrical expression and musical structure after generation.
The creation endpoint provides an instrumental switch, which you can use to set an instrumental task, then describe the instruments, mood, and style with prompts. An instrumental setting does not mean specifying an exact song length or receiving stem files; upon delivery, you should read the actual audio and duration, and perform post-processing according to the intended use.
After completion, read audio_url from data, and save the song ID and status. When using async=true, first obtain task_id, then retrieve the result through task queries; you can also set callback_url to receive completion notifications. Task submission success and audio completion should be presented separately in the application.
Choose a development approach or application suited to this model and follow its setup guide.
↗Python SDK Integration Guide
SDK installation, authentication and examples↗TypeScript SDK Integration Guide
SDK installation, authentication and examples
↗Pricing
$0.0552· per generation
You pay$13,584for 142,800 credits · $0.0951 / credits46% OFF
You're billed in credits. Pick a top-up tier to see its per-credit rate — the larger the top-up, the cheaper each credit.
This is only a basic call example. See the full docs for more parameters and advanced usage.
View Docs