What is Suno Inspiration?
Suno Inspiration is an endpoint that creates a new song with your earlier Suno clips as a style guide. You start with clips from previous Suno jobs, pass one to four of their IDs to suno-ai/inspiration in clip_ids, and the model writes a fresh track that sounds like it belongs with them. The references set the direction. Your inputs set the content: tags for genre and mood, prompt for custom lyrics, title for the song name, negative_tags for styles to avoid, and make_instrumental for a track without vocals. You also choose the model version with mv. Each task returns a pair of new clips in the same shape as a normal music generation. It is the right call when you want "more like these" rather than a direct blend of two tracks.
Key features of Suno Inspiration
Four features cover what you control on an inspiration request.
One to four reference clips
Pass a single clip to match one song, or up to four to describe a sound that sits across several tracks. The clip_ids array accepts any clips from your finished Suno jobs.
New composition, not a copy
The references guide style and feel. The melody and lyrics come out new, shaped by your tags, prompt, and title.
Instrumental mode and style exclusions
Set make_instrumental to true for a track without vocals, and use negative_tags to keep genres you do not want out of the result.
Three v6 model versions
Choose chirp-hawk (v6), chirp-hawk-wild (v6 tuned for experimental output), or chirp-goose (v6 Mini, faster and lighter) with the mv field.
Best for
Album consistency
Feed the first few tracks of a project as references so new songs keep the same sound.
Variations on a strong track
Use one clip that worked well as the only reference and generate new songs in its style.
Brand background music
Keep a set of approved clips as references and generate fresh themed tracks that match them.
Mood boards in audio
Pick up to four clips that capture a feeling and let the model write a song around them.
Playlist fillers
Generate tracks that sit naturally between existing songs in a playlist.
Use cases
A game studio can store four approved soundtrack clips as references and generate new level music on demand, with make_instrumental set to true. A creator app can show a "make more like this" button on any song in the user library that sends that clip as the only reference. An agency can keep one reference set per client and generate fresh tracks for each campaign that still sound on brand. For a direct blend of two tracks use the Suno Mashup API, and generate the reference clips themselves with Suno music generation.
API examples
The Suno remix docs list every parameter for inspiration and its sibling endpoints.
curl -X POST https://api.unifically.com/v1/tasks \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"model": "suno-ai/inspiration",
"input": {
"mv": "chirp-hawk",
"clip_ids": ["1901ec96-72c6-4c26-8e56-91028ca6070d"],
"tags": "rainy night, ambient",
"make_instrumental": true
}
}'
You get a task_id back right away. Poll GET /v1/tasks/<task_id> for the two finished clips, or pass a callback_url outside input to have the result posted to your server.
FAQs
People also ask
It writes a new song using one to four of your earlier Suno clips as references. The references shape the sound and feel, while your tags, prompt lyrics, and title shape the new track. Each request returns a pair of clips.
Between one and four clip IDs in the clip_ids array. A single reference works fine when you want one song to set the mood.
Mashup (suno-ai/mashup) takes exactly two clips and blends them into one song. Inspiration takes one to four clips and uses them as a guide for a new composition, so the result follows their style without being a blend of the source tracks.
No. The references steer style and feel, and the output is a new composition with your own tags and lyrics.
Yes. Set make_instrumental to true. This works well for background music that should match a reference track without vocals.
Yes. Each request generates new audio and is billed per request. Check the pricing page for the live rate. Failed tasks are never charged.
