Skip to main content
Music generation creates original compositions from a text description. Describe the genre, mood, instruments, and tempo and the agent produces an audio file.

​
How to generate music

1

Describe the music you want

Type a prompt describing the style, mood, and characteristics of the music.

Lo-fi background music.

Upbeat corporate jingle.

Cinematic orchestral score.

2

Set the duration (optional)

Mention the duration in your prompt. If you do not, the agent chooses a reasonable length.
Duration rangeNotes
3 secondsMinimum
600 seconds (10 minutes)Maximum
3

Review and download

The music file appears in the chat with a waveform player. Listen to it and download as MP3.
Be specific about instruments, tempo, and mood. “Ambient electronic with soft pads and a slow tempo” produces better results than “some chill music.”

​
What you cannot do

  • You cannot generate music longer than 10 minutes in a single request.
  • You cannot add lyrics to generated music. Music generation is instrumental only.
  • You cannot layer music over speech in a single operation. Generate them separately and combine in an external editor.
  • You cannot edit or remix a generated track. Generate a new version with an updated prompt.

​
Next steps

Sound Effects

Generate short sound effects from a text description.

Text-to-Speech

Generate speech to pair with your background music.