Introduction
The Suno v6 cover endpoint takes a finished song and performs it again in a style you describe. The melody and, by default, the words come from the recording you send in audio_url; the arrangement, instruments, production and singing voice follow prompt. It is the API counterpart of making a cover version: same tune, new character. Each request returns two complete versions, so you can choose between two takes on the same idea.
Key features & highlights
The tune survives the restyle
We covered three songs made on our own platform, an 80s synth-pop track, an indie-pop song and a lo-fi hip-hop piece, into acoustic folk, lounge jazz and an orchestral ballad. After separating the vocals of each result and aligning them with the original vocal, 67–91% of the notes landed within half a semitone of the original pitch. Sixteen of the eighteen versions stayed in the original key or its relative key, which keeps a cover usable alongside the source.
A new arrangement, not a filter
The cover is regenerated from scratch rather than processed: tempo feel, instrumentation and song length can all change. A 73-second song came back between 78 and 116 seconds long because the new arrangement added intros and instrumental passages. That freedom is what makes a genre change convincing, and it is also why the length is a guide rather than a guarantee.
Fine controls stay available
The style controls of Suno’s generator carry over: style_weight, weirdness, variety, vocal_gender, a negative style list and saved personas. audio_weight balances the uploaded recording against the new style; in our measurements its effect was gentle and varied from song to song.
Request parameters
| Parameter | Required | Description |
|---|---|---|
audio_url | Yes | The song to cover, as MP3 or WAV of up to 8 minutes. Songs made with Suno work; so do your own recordings, as long as you hold the rights. at least 1 characters |
prompt | Yes | Describe the destination sound in up to 1,000 characters: genre, tempo feel, instruments and the singer’s voice. Leave the tune out; it comes from the recording. at least 1 characters · at most 1000 characters |
lyrics | No | Paste the words to sing, normally the original lyrics with any lines you want changed. Bracketed section tags guide the structure and are not sung. at least 1 characters · at most 5000 characters |
audio_weight | No | Balance between the uploaded recording and the new style, 0–1 in steps of 0.01. In our tests the effect was small and not the same on every song. |
instrumental | No | Turn on for a cover without vocals; lyrics and |
title | No | Metadata title for the cover, up to 80 characters; it is not sung. at least 1 characters · at most 80 characters |
duration_seconds | No | Target length from 10 to 360 seconds. Suno treats it as a guide: 60 returned versions of 65 and 78 seconds. |
negative_prompt | No | Up to 500 characters of styles or sounds to avoid in the new arrangement. at least 1 characters · at most 500 characters |
vocal_gender | No | Raises the chance of a female or male lead voice; it is a preference, not a guarantee. |
style_weight | No | How strictly the cover follows the style in prompt, from 0 to 1. |
weirdness | No | Room for unusual choices in the arrangement, from 0 (conventional) to 1 (adventurous). |
variety | No | Spread between the two returned versions: 0 off, 1 normal, up to 4. |
persona_id | No | ID of a saved Suno persona whose voice or style should carry into the cover; send |
persona_model | No |
|
Pricing
$0.072 / request
One flat price per cover request, whatever the song length; each request returns two cover versions.
Use cases
Genre versions of your own song
Release an acoustic or orchestral take next to the original, or test which genre suits a melody before booking studio time. Keep the original lyrics in lyrics so the words stay identical and only the sound changes, then compare the two returned versions side by side.
Rework a demo into a fuller production
Send a rough demo you recorded and describe the production you have in mind. The cover keeps the tune you wrote and builds a full arrangement around it, which is quicker than describing the melody in text and hoping a generator finds it.
Prompt guide & pro tips
Describe the destination, not the source
Write prompt as if briefing a session band: genre, tempo feel, the two or three instruments that define the sound, and the voice. "Acoustic unplugged folk, fingerpicked steel-string guitar, upright bass, brushed drums, warm intimate male vocal" produced a convincing folk version of a synth-pop song in our test. Do not describe the melody; it comes from the recording.
Keep the words in lyrics
Paste the original lyrics with section tags such as [Verse] and [Chorus]. Edit individual lines to change what is sung while keeping the structure. Turn on instrumental for a version without vocals; lyrics and vocal_gender are then rejected.
Check both versions and the length
Listen to both takes before choosing: they share the brief but differ in arrangement and length. If you need a shorter cover, set duration_seconds and expect the result to land near it rather than on it. Use recordings you have the rights to; Suno refuses sources it identifies as copyrighted.
API quick start
# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
-H "Authorization: Bearer $SPICY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
-d '{
"model": "suno/v6/audio-cover",
"input": {
"audio_url": "https://example.com/reference.mp3",
"prompt": "Calm ambient pads with a slow rising motif"
}
}'
# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
-H "Authorization: Bearer $SPICY_API_KEY"
See the API reference for authentication, the complete schema, polling, and error handling.


