Voices

List available voices from audio providers. Useful for building voice selectors and previewing options before generating speech.

List Voices

use Atlasphp\Atlas\Atlas;

$voices = Atlas::provider('openai')->voices();

foreach ($voices->voices as $voiceId) {
    echo "$voiceId\n";
}

Using Voices

Pass a voice ID to any audio generation request:

$response = Atlas::audio('openai', 'tts-1')
    ->instructions('Hello, welcome to Atlas!')
    ->withVoice('nova')
    ->asAudio();

Building a Voice Selector

// In a controller
public function voices(string $provider)
{
    $voices = Atlas::provider($provider)->voices();

    return response()->json($voices);
}

// In a request
$response = Atlas::audio($request->provider, $request->model)
    ->instructions($request->text)
    ->withVoice($request->voice)
    ->asAudio();

ElevenLabs Voices

ElevenLabs provides a large library of voices including community-created options:

$voices = Atlas::provider('elevenlabs')->voices();

foreach ($voices->voices as $voiceId) {
    echo "$voiceId\n";
    // "21m00Tcm4TlvDq8ikWAM"
    // "AZnzlk1XvdvUeBnXmlld"
    // ...
}

Use the voice ID from the listing:

$response = Atlas::audio('elevenlabs', 'eleven_multilingual_v2')
    ->instructions('Welcome to Atlas.')
    ->withVoice('21m00Tcm4TlvDq8ikWAM')
    ->asAudio();

Caching

Voice listings are cached by default to minimize API calls:

ATLAS_CACHE_VOICES_TTL=3600   # 1 hour (default)

Set TTL to 0 to disable caching.

Supported Providers

Provider	Voices	Source
OpenAI	alloy, ash, ballad, coral, echo, fable, onyx, nova, sage, shimmer	Hardcoded
xAI	Fetched from API (/v1/tts/voices)	Live API
ElevenLabs	Large voice library + community voices	Live API

API Reference

Method	Returns	Description
voices()	VoiceList	List available voices from the provider