elevenlabs-tts

# Eleven Labs Text-to-Speech

Generate high-quality audio from text using the Eleven Labs API. Ideal for podcast-style summaries, narration, and voice-overs.

## Quick Start

```bash
# Generate audio with default voice (Rachel)
uv run scripts/generate_audio.py --text "Hello, this is a test."

# Generate from a text file
uv run scripts/generate_audio.py --file report.txt --output ~/Desktop/report.mp3

# Use a specific voice
uv run scripts/generate_audio.py --text "Breaking news..." --voice Josh

# List available voices
uv run scripts/list_voices.py
```

## Scripts

### generate_audio.py

Main script for TTS generation.

**Arguments:**
| Argument | Description | Default |
|----------|-------------|---------|
| `--text` | Text content to convert | - |
| `--file` | Path to text file | - |
| `--voice` | Voice name or ID | Rachel |
| `--model` | Model ID | eleven_multilingual_v2 |
| `--output` | Output file path | ~/Desktop/audio_TIMESTAMP.mp3 |

**Features:**
- Auto-chunks text >10k characters at sentence boundaries
- Concatenates chunks into single MP3
- Returns JSON with metadata (file path, voice, model, char count, chunks)

### list_voices.py

Fetch and display available voices.

**Arguments:**
| Argument | Description | Default |
|----------|-------------|---------|
| `--all` | Show all voices (not just premade) | false |
| `--json` | Output as JSON | false |

## Curated Voices

Six voices selected for variety (see `references/voices.md` for details):

| Voice | Style | Best For |
|-------|-------|----------|
| **Rachel** | Calm, clear | Narration, podcasts (default) |
| **Bella** | Soft, gentle | Storytelling, meditation |
| **Elli** | Young, expressive | Casual content, tutorials |
| **Josh** | Deep, authoritative | News, professional content |
| **Adam** | Middle-aged, clear | Business, documentaries |
| **Antoni** | Warm, versatile | General purpose |

Use `list_voices.py` to discover additional voices.

## Models

| Model | ID | Best For | Char Limit |
|-------