---
name: elevenlabs-tts
description: Convert text to speech using ElevenLabs voice AI. Supports 70+ languages, multiple voices, and models from highest quality to ultra-low latency.
---

# ElevenLabs Text-to-Speech

Generate natural-sounding audio from text using ElevenLabs voice AI.

## Prerequisites

- **ELEVENLABS_API_KEY** environment variable set

## Commands

### Generate speech

```bash
/root/clawd/skills/elevenlabs-tts/scripts/tts.sh "Your text here"
/root/clawd/skills/elevenlabs-tts/scripts/tts.sh "Your text here" output.mp3
/root/clawd/skills/elevenlabs-tts/scripts/tts.sh "Your text here" output.mp3 <voice_id> <model_id>
```

Outputs the audio file path (saved to `/tmp/elevenlabs_tts_output.mp3` by default).

### List available voices

```bash
/root/clawd/skills/elevenlabs-tts/scripts/list-voices.sh
```

Returns JSON with voice IDs, names, and labels.

## Popular Voices

| Voice ID | Name | Style |
|----------|------|-------|
| `JBFqnCBsd6RMkjVDRZzb` | George | Multilingual, warm |
| `21m00Tcm4TlvDq8ikWAM` | Rachel | English, calm |
| `AZnzlk1XvdvUeBnXmlld` | Domi | English, confident |
| `EXAVITQu4vr4xnSDxMaL` | Bella | English, soft |

## Models

| Model ID | Description |
|----------|-------------|
| `eleven_multilingual_v2` | Best quality, 70+ languages (default) |
| `eleven_flash_v2_5` | Ultra-low latency (~75ms) |
| `eleven_turbo_v2_5` | Low latency, English optimized |

## When to use

- User asks to read text aloud or generate a voiceover
- Creating audio content from written text
- Accessibility: reading documents or messages as speech
- Generating audio for videos or presentations