What is Suno Add Instrumental?
Suno Add Instrumental is the reverse of Add Vocals: you bring the voice or melody, and it writes the band. The input must be audio you uploaded, such as an a cappella take or a voice memo of a melody, referenced by clip_id or sent as audio_url. Clips produced by Suno generation are not accepted. You describe the backing either as a short brief in gpt_description_prompt ("Hip-hop beats with 808 bass") or as a more specific direction in prompt ("Rock guitar and drums with heavy bass"). Optional start_s and end_s values limit the arrangement to a section of the recording. The job runs async and returns the track with the new instrumental added.
Key features of Suno Add Instrumental
Four parts of the request matter most.
Arrangements under your own takes
Start from an uploaded vocal, melody or rough recording, referenced by clip_id or audio_url. Your performance stays the starting point and the music is written around it.
Short brief or detailed direction
Use gpt_description_prompt for a one-line genre brief, or prompt when you want to name specific instruments, feel and energy.
Section control in seconds
start_s and end_s limit the backing to one part of the recording, so you can arrange the chorus first or score only the section a video needs.
Your own production style
Choose chirp-hawk, chirp-hawk-wild or chirp-goose, or set mv to chirp-custom with a custom_model of 6 to 24 reference tracks so the backing follows the sound of your past work. A title labels the output.
Best for
Singers without a band
Record a dry vocal and get a full backing arrangement to sing over or share as a demo.
Melody sketches
Turn a hummed or whistled idea into an arranged demo before taking it into a DAW.
Trying different genres under one vocal
Send the same take with a rock brief, then an electronic one, and compare the results.
Spoken word with music
Add a short scored bed under a poem, intro or voice-over, limited with start_s and end_s.
Practice tracks
Give vocal students an arranged backing built around their own recorded take.
Use cases
A singing app can record the user's a cappella take, upload it, and offer genre buttons that each send a different brief to Add Instrumental. A podcast tool can add a short instrumental bed under the host's recorded intro by setting end_s to the length of the intro. A songwriting workflow can run the same melody recording through three Custom mode prompts and save the arrangements as alternate demos. When you have the opposite problem, an instrumental that needs a singer, use the Suno Add Vocals API.
API examples
Parameter docs: docs.unifically.com/models/audio/suno/add-instrumental.
curl -X POST https://api.unifically.com/v1/tasks \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_API_KEY" \
-d '{
"model": "suno-ai/add-instrumental",
"input": {
"mv": "chirp-hawk",
"clip_id": "ec0be483-2de2-4abc-a5ce-3d8258f52ed4",
"gpt_description_prompt": "Hip-hop beats with 808 bass",
"start_s": 0,
"end_s": 60
}
}'
The response carries a task_id. Poll GET /v1/tasks/<task_id> for the result, or add a callback_url to the request to be notified when the arrangement is done.
FAQs
People also ask
It writes music under audio you supply. Upload a dry vocal, a hummed melody or a rough take, describe the backing you want, and the job returns the audio with a generated instrumental arrangement around it.
No. Add Instrumental only works with uploaded audio. Reference your upload by clip_id or pass the file as audio_url. To restyle a Suno-generated song, use Cover instead.
gpt_description_prompt is the Simple mode brief, a short line like "Hip-hop beats with 808 bass". prompt is the Custom mode field for more specific direction, such as "Rock guitar and drums with heavy bass". Send whichever fits how much control you need.
Yes. Set start_s and end_s in seconds to limit the section that gets the instrumental, for example 0 to 60 for the first minute.
chirp-hawk (v6), chirp-hawk-wild (v6 Wild) or chirp-goose (v6 Mini). For a backing in the style of your own productions, set mv to chirp-custom and pass a custom_model with 6 to 24 reference audio URLs.
