Riffusion AI Music Generation API#
Riffusion is a cutting-edge AI music creation platform similar to Suno or Udio, now offering one of the most efficient and high-quality music generation on the market.π Why Riffusion?#
π΅ Near Suno v4.5 Quality: Delivers studio-level music quality, including vocals and instrumentals.
π Unique Swap Capabilities: Swap vocals or sound style in any music clip β a feature unmatched by other platforms.
β‘ Fast Generation: Produces 3β4 minute songs in under 60 seconds.
πΈ Affordable & Stable: Low cost per track and reliable performance at scale.
π§ Flexible Features: Supports lyrics-based, prompt-based, instrumental, and remix workflows.
π Commercial License: All generated content is licensed for commercial use.
π Credit Consumption#
Action Type | Credit Usage |
---|
upload tasks | 2 credits |
create tasks | 10 credits |
π‘ API Parameters#
Field | Type | Required | Description |
---|
custom_mode | boolean | β
Yes | Set to true to use lyrics-based generation, false for prompt-based. |
lyrics | string | β No | Lyrics input (used only when custom_mode is true ). |
title | string | β No | Title of the song (metadata, optional, only with custom_mode ). |
make_instrumental | boolean | β
Yes | Whether to generate instrumental-only music. |
prompt | string | β No | Prompt or theme for the song (used only when custom_mode is false ). |
sound | string | β No | Style and sound design (genre, voice style, etc., only with custom_mode ). |
mv | enum | β
Yes | Model version. Options: "FUZZ-1.1 Pro" , "FUZZ-1.1" , "FUZZ-1.0 Pro" , "FUZZ-1.0" . |
task_type | string | β No | Task type: create_music , extend_music , cover_music , extend_upload_music , replace_music , replace_upload_music , swap_music_sound , swap_upload_music_sound , swap_music_vocals , swap_upload_music_vocals . |
continue_clip_id | string | β No | The clip ID of the original song to operate on (uploaded clip or previously generated one). |
continue_at | number | β No | Starting second for extension (extend task types only). |
replace_start_time | number | β No | Start time (in seconds) of the segment to replace. |
replace_end_time | number | β No | End time (in seconds) of the segment to replace. |
vocal_variation | number | β No | Degree of variation in vocals or instruments (used only in cover_music , swap_music_vocals , and swap_music_sound task types). |
π Response & Polling Notes#
Riffusion generates full tracks quickly (typically under 1 minute). There is no intermediate status, so when a generation is in progress, the API will return:{
"type": "processing",
"message": "processing generation, please try again later."
}
βΉοΈ Additional Notes#
β± Recommended polling interval: every 10 seconds.
β οΈ When using an uploaded clip_id
, ensure the task_type
is the correct upload variant (e.g., extend_upload_music
, swap_upload_music_sound
, etc.).
π There is no intermediate progress β keep polling until the final result is returned.
Modified atΒ 2025-07-22 20:36:08