Music Generation

Generate original, royalty-free music from text descriptions using AI.

Overview

The Music Generation API creates original compositions from natural language prompts. Control genre, tempo, duration, and output format to produce royalty-free tracks ready for any project. Stems separation lets you isolate individual instruments for remixing.

Text-to-music generation

Genre control

Tempo & duration control

Multiple formats

Royalty-free output

Stems separation

Quickstart

Generate a music track from a text description in just a few lines.

from nur import NurClient
client = NurClient()
# Generate a track from a prompt
track = client.music.generate(
    prompt="Upbeat lo-fi hip hop beat with jazzy piano chords and vinyl crackle",
    duration=30,
    genre="lo-fi",
    tempo=85
)
print(f"Track ID: {track.id}")
print(f"Audio URL: {track.audio_url}")

Generate Music

POST/v1/music/generate

Generate an original music track from a natural language description. The model interprets your prompt to determine instrumentation, mood, and structure. Optional parameters let you constrain genre, tempo, duration, and output format.

Parameter	Type	Description
promptREQUIRED	string	Natural language description of the desired music.
duration	integer	Track length in seconds (5-300). Default: 30.
genre	string	Target genre (e.g., lo-fi, electronic, classical, rock, ambient).
tempo	integer	Tempo in BPM (40-220). Auto-detected from prompt if omitted.
output_format	string	Audio format: mp3, wav, flac, ogg. Default: mp3.

# Generate a cinematic orchestral piece
track = client.music.generate(
    prompt="Epic cinematic orchestral soundtrack with sweeping strings and brass",
    duration=60,
    genre="classical",
    tempo=100,
    output_format="wav"
)
print(f"Track ID: {track.id}")
print(f"Status: {track.status}")
print(f"Audio URL: {track.audio_url}")
print(f"Genre: {track.genre}, Tempo: {track.tempo} BPM")
print(f"Duration: {track.duration}s")

Response

{
  "id": "trk_m8k2v9x1q4w7",
  "status": "completed",
  "audio_url": "https://cdn.nur.ai/music/trk_m8k2v9x1q4w7.wav",
  "duration": 60,
  "genre": "classical",
  "tempo": 100,
  "prompt": "Epic cinematic orchestral soundtrack with sweeping strings and brass",
  "created_at": "2025-01-15T14:20:00Z"
}

Get Track

GET/v1/music/{track_id}

Retrieve a generated music track by its ID. The response includes download URLs, waveform data for visualization, and separated stems for remixing.

Parameter	Type	Description
track_idREQUIRED	string	The unique identifier of the music track.

# Retrieve a track with stems
track = client.music.get("trk_m8k2v9x1q4w7")
print(f"Track: {track.id}")
print(f"Download: {track.download_url}")
print(f"Waveform: {track.waveform_url}")
# Access individual stems for remixing
print(f"Vocals: {track.stems.vocals}")
print(f"Drums: {track.stems.drums}")
print(f"Bass: {track.stems.bass}")
print(f"Other: {track.stems.other}")

Response

{
  "id": "trk_m8k2v9x1q4w7",
  "status": "completed",
  "audio_url": "https://cdn.nur.ai/music/trk_m8k2v9x1q4w7.wav",
  "download_url": "https://cdn.nur.ai/music/trk_m8k2v9x1q4w7/download.wav",
  "waveform_url": "https://cdn.nur.ai/music/trk_m8k2v9x1q4w7/waveform.json",
  "duration": 60,
  "genre": "classical",
  "tempo": 100,
  "prompt": "Epic cinematic orchestral soundtrack with sweeping strings and brass",
  "created_at": "2025-01-15T14:20:00Z",
  "stems": {
    "vocals": "https://cdn.nur.ai/music/trk_m8k2v9x1q4w7/stems/vocals.wav",
    "drums": "https://cdn.nur.ai/music/trk_m8k2v9x1q4w7/stems/drums.wav",
    "bass": "https://cdn.nur.ai/music/trk_m8k2v9x1q4w7/stems/bass.wav",
    "other": "https://cdn.nur.ai/music/trk_m8k2v9x1q4w7/stems/other.wav"
  }
}

Response Objects

Reference for the objects returned by the Music Generation endpoints.

MusicTrack (Generate)

Field	Type	Description
id	string	Unique identifier for the music track.
status	string	One of: pending, processing, completed, failed.
audio_url	string	URL to stream the generated audio.
duration	number	Track length in seconds.
genre	string	Detected or specified genre of the track.
tempo	integer	Tempo of the track in BPM.
prompt	string	The original text prompt used for generation.
created_at	string	ISO 8601 timestamp of track creation.

MusicTrack (Get - Full)

Field	Type	Description
id	string	Unique identifier for the music track.
status	string	One of: pending, processing, completed, failed.
audio_url	string	URL to stream the generated audio.
download_url	string	Direct download URL for the audio file.
waveform_url	string	URL to the waveform data JSON for visualization.
duration	number	Track length in seconds.
genre	string	Detected or specified genre of the track.
tempo	integer	Tempo of the track in BPM.
prompt	string	The original text prompt used for generation.
created_at	string	ISO 8601 timestamp of track creation.
stems	object	Separated stems with URLs for vocals, drums, bass, and other.

Best Practices

Write descriptive prompts

Include details about mood, instruments, and style. "Chill lo-fi beat with jazzy Rhodes piano, soft drums, and rain ambiance" produces far better results than "lo-fi music".

Use WAV for production, MP3 for previews

Generate preview tracks in MP3 for faster iteration, then re-generate your final selection in WAV or FLAC for lossless quality in production.

Leverage stems for custom mixes

Use the stems separation feature to isolate drums, bass, vocals, and other instruments. This lets you remix, adjust levels, or combine elements from multiple generations.

Set tempo explicitly for rhythmic content

If your project requires precise BPM (e.g., syncing to video), always specify the tempo parameter rather than relying on auto-detection from the prompt.