Suno API: Python SDK for Music and Audio Generation

A focused Python SDK for Suno music and audio workflows through MuAPI. Create songs, remix or extend tracks, generate lyrics and sound effects, add vocals or instrumentals, make mashups, convert tracks to WAV, and run Suno's two-stage custom voice flow with one API key.

Related Projects

• Suno on MuAPI — model landing page for Suno's music and audio capabilities.
• Suno create music playground — try the primary workflow in the browser.
• MuAPI music and speech docs — audio endpoint and workflow reference.
• suno-comfyui — focused ComfyUI nodes for Suno create, remix, extend, and utility workflows.
• minimax-music-3-api — sibling Python SDK for MiniMax Music 3.0 text-to-music generation.
• awesome-minimax-music-3-prompts — curated music prompts and a practical lyrics-formatting guide.
• minimax-music-3-comfyui — ComfyUI nodes for a related text-to-music model.
• midjourney-api — sibling Python SDK for Midjourney image-generation workflows.
• Open-Generative-AI — open-source studio for running generative image, video, and audio workflows.
• Generative-Media-Skills — agent-ready skills for building generative-media pipelines.
• muapi-cli — CLI and MCP access to the same MuAPI model catalog.

Features

• Create music with Suno V3.5 through V5.5 model options
• Remix and extend existing audio URLs
• Generate lyrics, sound effects, and expanded music-style descriptions
• Add vocals, add instrumentals, or create mashups from multiple tracks
• Convert a completed track to lossless WAV
• Two-stage custom voice cloning with phrase verification
• Async polling, webhook URLs, output URL extraction, and injectable HTTP sessions

Installation

pip install suno-api

Or install the latest source:

git clone https://github.com/Anil-matcha/suno-api.git
cd suno-api
pip install -e .

Set your MuAPI key:

export MUAPI_API_KEY=your_muapi_api_key

Create a key from the MuAPI access-key page.

Quick start

from suno_api import SunoAPI

api = SunoAPI()

task = api.create_music(
    style="Hip-Hop",
    prompt="A confident song about winning with a memorable hook",
    instrumental=False,
    model="V5",
)

result = api.wait_for_completion(task["request_id"])
for audio_url in api.extract_audio_urls(result):
    print(audio_url)

For a blocking helper, use api.create_music_and_wait(style="Hip-Hop", prompt="...", instrumental=False).

Core workflows

Remix and extend

remix = api.remix_music(
    audio_url="https://example.com/source.mp3",
    style="warm lo-fi jazz with brushed drums",
    prompt="Keep the melody recognizable but make the mix intimate",
)

extension = api.extend_music(
    audio_url="https://example.com/source.mp3",
    style="cinematic electronic pop",
    continue_at=42,
)

Lyrics, sound effects, and style expansion

lyrics = api.generate_lyrics("A hopeful song about rebuilding after a storm")
sound = api.generate_sounds("Distant thunder rolling across a mountain valley")
expanded_style = api.boost_music_style("dark electronic pop")

Vocals, instrumentals, and mashups

vocals = api.add_vocals(
    prompt="A confident lead vocal with a wide chorus",
    title="After the Rain",
    style="cinematic pop, uplifting, modern drums",
    audio_url="https://example.com/instrumental.mp3",
)

instrumental = api.add_instrumental(
    title="After the Rain Backing Track",
    tags="cinematic pop, modern drums, warm piano",
)

mashup = api.generate_mashup(
    ["https://example.com/intro.mp3", "https://example.com/chorus.mp3"],
    style="indie electronic",
    instrumental=True,
)

Custom voice cloning

Voice cloning uses a required two-stage verification flow. The first request returns a fresh phrase; record yourself reading that phrase, upload the recording, then confirm it:

stage_one = api.voice_clone(
    "https://example.com/clean-voice-sample.wav",
    voice_name="My Singing Voice",
    language="en",
)

# Poll until the result contains stage="awaiting_phrase" and a phrase.
phrase_state = api.get_result(stage_one["request_id"])
print(phrase_state["phrase"])

stage_two = api.confirm_voice_clone(
    stage_one["request_id"],
    "https://example.com/verification-phrase.wav",
)
completed = api.wait_for_completion(stage_two["request_id"])
print(completed.get("voice_id"))

The voice library helpers are list_voices(), check_voice(voice_db_id), refresh_voice(voice_db_id), and delete_voice(voice_db_id). Only clone voices you have permission to use.

API endpoints

Workflow	Endpoint
Create music	POST /api/v1/suno-create-music
Remix music	POST /api/v1/suno-remix-music
Extend music	POST /api/v1/suno-extend-music
Convert to WAV	POST /api/v1/suno-convert-to-wav
Generate sounds	POST /api/v1/suno-generate-sounds
Generate lyrics	POST /api/v1/suno-generate-lyrics
Boost music style	POST /api/v1/suno-boost-music-style
Add vocals	POST /api/v1/suno-add-vocals
Generate mashup	POST /api/v1/suno-generate-mashup
Add instrumental	POST /api/v1/suno-add-instrumental
Start voice clone	POST /api/v1/suno-voice-clone
Confirm voice clone	POST /api/v1/suno-voice-clone/{request_id}/confirm
Poll task	GET /api/v1/predictions/{request_id}/result

All requests use the x-api-key header. Inputs that reference audio must be publicly reachable URLs. Every generation returns a request_id immediately; use polling or provide webhook_url.

Raw request shape

curl --location --request POST "https://api.muapi.ai/api/v1/suno-create-music" \
  --header "x-api-key: $MUAPI_API_KEY" \
  --header "Content-Type: application/json" \
  --data-raw '{
    "style": "Hip-Hop",
    "prompt": "A confident song about winning with a memorable hook",
    "model": "V5",
    "instrumental": false,
    "webhook_url": "https://example.com/webhook"
  }'

For a completed result, extract_audio_urls(result) returns the hosted output URLs. extract_audio_ids(result) returns the track IDs needed by convert_to_wav().

Development

python -m unittest discover -s tests -v

The tests mock HTTP calls and never spend credits.

License

MIT — see LICENSE.