> ## Documentation Index
> Fetch the complete documentation index at: https://docs.melius.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Audio models

> Audio-generation models available on audio nodes.

Audio nodes generate sound from a prompt — speech, music, or sound effects — and can also transcribe speech or modify existing voice clips.

## Speech (text-to-speech)

* **ElevenLabs v3** — highest-quality, most expressive voice model.
* **ElevenLabs Multilingual v2** — broad language coverage.
* **ElevenLabs Turbo v2.5** — fastest, lowest cost.
* **Seed Audio 1.0** — speech generation from ByteDance.

## Sound effects

* **ElevenLabs Sound Effects** — short non-musical clips (1–22 seconds) from a text prompt. Use for foley, ambience, and quick sound cues.

## Music

* **ElevenLabs Music** — instrumental tracks (10–300 seconds) from a text prompt.

## Transcription (speech-to-text)

* **ElevenLabs Scribe v2** — transcribes spoken audio into text.

## Voice change (audio-to-audio)

* **ElevenLabs Voice Changer** — re-voices an existing audio clip with a different speaker while preserving timing and inflection.
* **ElevenLabs Voice Isolator** — separates speech from background noise in an existing audio clip.

Use [`model_list`](/mcp/tools#models) for the live catalog and exact variant IDs.