# ElevenLabs v3 - Core.Today AI API

> ElevenLabs' most expressive text-to-speech model. 26 premium voices, inline audio tags like [laughs] and [whispers], fine-grained style and stability controls, and 70+ language support.

- **Provider**: ElevenLabs
- **Model ID**: elevenlabs/v3
- **Category**: Audio & TTS
- **Credits**: 237 per 1000 characters (0.225 credits/char, billed per character)
- **Speed**: Medium
- **Quality**: Ultra

## Features
- Most expressive ElevenLabs model
- Inline audio tags: [laughs], [whispers], [sighs], [excited]
- 26 premium preset voices
- Style, stability, and similarity controls
- 70+ language support

## Use Cases
- Character dialogue for games and animation
- Dramatic audiobook narration
- Emotional storytelling content
- Premium video voiceovers
- Podcast intros and segments

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **prompt**: string - The text to convert to speech. Supports audio tags like [laughs], [whispers], [sighs]

### Optional
- **voice**: string (default: Rachel) - Voice choice for speech generation
  Options: Rachel, Drew, Clyde, Paul, Aria, Domi, Dave, Roger, Fin, Sarah, James, Jane, Juniper, Arabella, Hope, Bradford, Reginald, Gaming, Austin, Kuon, Blondie, Priyanka, Alexandra, Monika, Mark, Grimblewood
- **stability**: number (default: 0.5) - Voice stability (0.0-1.0). Lower is more variable and expressive
- **similarity_boost**: number (default: 0.75) - How closely the output follows the original voice (0.0-1.0)
- **style**: number (default: 0) - Style exaggeration (0.0-1.0)
- **speed**: number (default: 1) - Speed of speech (0.7-1.2)
- **previous_text**: string - Previous text for prosody context across segments
- **next_text**: string - Next text for prosody context across segments
- **language_code**: string (default: en) - Language of the text. Match it to your input for correct pronunciation
  Options: en, ko, ja, zh, de, hi, fr, pt, it, es, id, nl, tr, fil, pl, sv, bg, ro, ar, cs, el, fi, hr, ms, sk, da, ta, uk, ru, hu, no, vi

## Examples

### Expressive Storytelling
Dramatic narration with audio tags
```json
{
  "model": "elevenlabs/v3",
  "input": {
    "prompt": "[whispers] Something moved in the darkness. [pause] Then, out of nowhere... [excited] a tiny kitten jumped onto the table! [laughs]",
    "voice": "James",
    "stability": 0.4,
    "language_code": "en"
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- Use audio tags like [laughs], [whispers], [sighs] inline to direct the performance
- Lower stability (0.3-0.4) makes delivery more expressive; raise it for consistent narration
- Set language_code to match your text for correct pronunciation
- Use previous_text/next_text when splitting long scripts to keep prosody consistent
- For low-latency or high-volume use, consider Turbo v2.5 at half the price

## Documentation
https://replicate.com/elevenlabs/v3