What is Suno Add Vocals?
Suno Add Vocals takes an instrumental you already have and writes a sung vocal part on top of it. It is built for your own audio: the source must be a track you uploaded, not a clip that came out of Suno music generation. You reference that upload by clip_id, or pass the file as audio_url, then supply the words in gpt_description_prompt with section tags such as [Verse 1] and [Chorus]. Two optional fields, start_s and end_s, limit the vocal to a window in seconds, so an intro or outro can stay instrumental. The result is a new track with the vocal added, returned through the same async task flow as the rest of the Suno endpoints.
Key features of Suno Add Vocals
Four controls shape the vocal you get back.
Works on your own instrumentals
The input is audio you uploaded: a beat, a backing track or a score. Reference it with clip_id or send it as audio_url. Clips generated by Suno are not accepted here.
Your lyrics, your structure
Write the words in gpt_description_prompt with [Verse], [Chorus] and [Bridge] tags, or describe the vocal in plain language. Use the optional prompt field for extra direction.
Time window with start_s and end_s
Set where singing begins and ends in seconds. Useful for keeping a long intro clean or adding a vocal only to the chorus section of a track.
Model choice and custom vocal style
Pick chirp-hawk, chirp-hawk-wild or chirp-goose, or set mv to chirp-custom and pass a custom_model with 6 to 24 reference tracks so the vocal follows the style of your own songs. Add a title to label the result.
Best for
Producers with finished beats
Hear a topline and lyric over a beat before booking a session singer.
Testing lyric drafts
Run several versions of the lyrics over the same instrumental and compare them side by side.
Jingles and branded audio
Add a short sung hook over music you own or licensed, limited to the seconds where it should appear.
Sung sections in longer pieces
Use start_s and end_s to place a vocal inside one part of a score and keep the rest instrumental.
Demo songs from instrumentals
Turn an instrumental sketch into a full demo song with a guide vocal for pitching.
Use cases
A beat marketplace can let buyers preview a beat with a generated vocal by uploading the beat once and sending different lyrics against the same clip_id. A jingle tool can take a brand's licensed backing track, add a sung slogan between two timestamps, and export the result for ads. A songwriting app can pair this endpoint with the Suno Lyrics Generator API: generate two lyric options, pick one, and send it here to hear it sung over the user's own music.
API examples
See docs.unifically.com/models/audio/suno/add-vocals for the full parameter list.
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-vocals",
"input": {
"mv": "chirp-hawk",
"clip_id": "ec0be483-2de2-4abc-a5ce-3d8258f52ed4",
"gpt_description_prompt": "[Verse 1]\nNew vocal lyrics here",
"start_s": 10,
"end_s": 60
}
}'
Submitting returns a task_id. Poll GET /v1/tasks/<task_id> until the track is ready, or set a callback_url to receive it automatically.
FAQs
People also ask
It sings over music you supply. You send an instrumental you uploaded, the lyrics or a description of the vocal, and optionally a start and end time, and the job returns the track with a generated vocal on it.
No. Add Vocals only works with uploaded audio, not with clips produced by Suno generation. Reference your own file through its uploaded clip_id or pass it as audio_url.
Put them in gpt_description_prompt, using section tags like [Verse 1] and [Chorus] on their own lines. You can also describe the vocal instead of writing every line. A separate prompt field is available for extra direction.
They set the window, in seconds, where vocals are added. For example start_s 10 and end_s 60 leaves the first ten seconds as an instrumental intro and sings from there to the one-minute mark. Both are optional.
chirp-hawk (v6), chirp-hawk-wild (v6 Wild) and chirp-goose (v6 Mini). You can also set mv to chirp-custom with a custom_model of 6 to 24 reference tracks to follow the style of your own recordings.
