Guides

Audio generation

32 audio models cover text-to-speech, music, and sound effects. Each returns an audio URL on result.url.

Note

These snippets reuse an ai client — see Quickstart for the createClient setup.

Standard parameters#

The parameters most audio models share. Exact params vary per model — each model page lists its full schema.

Standard audio parameters
Name	Type	Description
`prompt`required	string	Text description of what to generate.
`voiceId`	string	Voice to speak with (see the model page for options).
`duration`	number	Length in seconds.

Text to speech#

Pass the text as prompt. Many voice models accept a voiceId to choose the speaker — the available voices are listed on each model page.

const speech = await ai.generate('eleven-v3', {
  prompt: 'Welcome to Picsart AI — let us build something amazing.',
  // Optional: pick a voice with voiceId (see the model page for options).
});

console.log(speech.url); // audio URL

Music#

Describe the track and set a duration in seconds.

const track = await ai.generate('elevenlabs-music-v2', {
  prompt: 'Upbeat lo-fi hip hop with a warm vinyl texture',
  duration: 30,
});

console.log(track.url);

Next steps#

Explore audio models →Voices, params, and examples.