Integration Guide

Suno submits music or lyric tasks first, then uses polling or webhooks to retrieve songs, lyrics, WAV files, and timing data.

Suno uses an asynchronous task model. The music generation endpoint only submits a task; it does not return the final audio synchronously. Store the returned task ID, then call /suno/fetch/{task_id} or receive completion through callback_url.

Scenario	Endpoint	Notes
Generate music	`POST /suno/submit/music`	Supports inspiration mode, custom lyrics, instrumental generation, continuation, and stem separation
Generate lyrics	`POST /suno/submit/lyrics`	Creates a lyric-generation task from a prompt
Fetch one task	`GET /suno/fetch/{task_id}`	Reads status, progress, songs, lyrics, and metadata
Batch fetch tasks	`POST /suno/fetch`	Refreshes multiple task IDs
Fetch song feed	`GET /suno/feed/{clip_id}`	Fetches song details by clip ID
Get WAV	`GET /suno/act/wav/{clip_id}`	Returns a WAV file URL
Get timing	`GET /suno/act/timing/{clip_id}`	Returns lyric and audio timing data

Base Paths

Default Suno-style endpoints:

https://{BASE_URL}/suno/submit/music
https://{BASE_URL}/suno/fetch

Compatible prefixes:

Format	Prefix
GoAmz	`{{BaseURL}}/suno/v1`
GoAmz v3.5	`{{BaseURL}}/suno/v1/mv-3.5`
SunoAPI	`{{BaseURL}}/suno`

Generate Music

custom_mode selects the generation mode. The default 0 is inspiration mode; 1 is custom lyrics mode. In inspiration mode prompt can be up to 200 characters; in custom mode it can be up to 3000 characters and is suitable for full lyrics.

curl --location 'https://api.modelsell.com/suno/submit/music' \
  --header 'Authorization: Bearer $MODELSELL_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{
    "callback_url": "https://example.com/webhooks/suno",
    "custom_mode": 1,
    "make_instrumental": 0,
    "prompt": "[Verse]\nWind moves without a sound\nShadows drift like a dream\n\n[Chorus]\nThe long season keeps its memory",
    "mv": "chirp-v4",
    "title": "Long Season",
    "tags": "Trance, bass drop, female vocal",
    "negative_tags": "dance pop"
  }'

The response can contain either one task ID or an array of task objects. Persist all returned task IDs and keep polling:

{
  "code": 200,
  "msg": "成功",
  "data": [
    {
      "task_id": "f9761a3e-b396-4060-a35b-a26bbce7ba9b",
      "status": 1,
      "title": "",
      "prompt": "美女"
    }
  ],
  "exec_time": 0.659346
}

Models And Key Fields

Field	Notes
`mv`	Model version. Supported values include `chirp-v4`, `chirp-v3-5`, `chirp-bluejay`, `chirp-auk`, and `chirp-fenix`; `chirp-fenix` maps to Suno 5.5 / latest
`custom_mode`	`0` for inspiration mode, `1` for custom lyrics mode
`make_instrumental`	Pass `1` to generate instrumental music
`prompt`	Up to 200 characters in inspiration mode, up to 3000 in custom mode; famous artist names may fail to generate
`tags`	Style tags, up to 120 characters
`negative_tags`	Styles or elements to exclude
`callback_url`	Webhook URL. Completed results are sent with POST
`continue_clip_id`	Previous official clip_id for continuation or stem separation; an existing task_id may also be accepted
`continue_at`	Continuation timestamp in seconds
`metadata.control_sliders`	Advanced controls such as `style_weight`, `weirdness_constraint`, and `audio_weight`

Poll Tasks And Read Results

The task response contains status and result data. On success, read data[].audio_url for MP3, data[].image_url for cover art, and data[].metadata for style and prompt information.

curl --location 'https://api.modelsell.com/suno/fetch/f9761a3e-b396-4060-a35b-a26bbce7ba9b' \
  --header 'Authorization: Bearer $MODELSELL_API_KEY'

import os
import time
import requests

base_url = "https://api.modelsell.com"
headers = {"Authorization": f"Bearer {os.environ['MODELSELL_API_KEY']}"}
task_id = "f9761a3e-b396-4060-a35b-a26bbce7ba9b"

while True:
    resp = requests.get(f"{base_url}/suno/fetch/{task_id}", headers=headers, timeout=30)
    resp.raise_for_status()
    payload = resp.json()
    task = payload.get("data", payload)

    if task.get("status") == "SUCCESS":
        for song in task.get("data", []):
            print(song.get("title"), song.get("audio_url"))
        break

    if task.get("status") == "FAILURE":
        raise RuntimeError(task.get("fail_reason") or "Suno task failed")

    time.sleep(5)

Lyrics, WAV, And Timing

Generate lyrics first when you want to review or edit lyrics before sending them to music generation:

curl --location 'https://api.modelsell.com/suno/submit/lyrics' \
  --header 'Authorization: Bearer $MODELSELL_API_KEY' \
  --header 'Content-Type: application/json' \
  --data '{"prompt":"dance"}'

After a music task succeeds, use the song clip ID to request WAV or timing data:

curl --location 'https://api.modelsell.com/suno/act/wav/90f72668-ba0d-4fa7-95e6-f71bebda88b7' \
  --header 'Authorization: Bearer $MODELSELL_API_KEY'

curl --location 'https://api.modelsell.com/suno/act/timing/90f72668-ba0d-4fa7-95e6-f71bebda88b7' \
  --header 'Authorization: Bearer $MODELSELL_API_KEY'

Integration Tips

Persist task_id, song clip_id, submit payload, status, audio URL, cover URL, and failure reason on your backend. Treat submitted, generating, succeeded, and failed as separate frontend states. For reliable delivery, poll or receive webhooks on the backend first, save results, then push updates to the frontend.