Skip to main content
The Voices API is a set of utility / discovery endpoints. They are read-only helpers and do not consume credits.

​
What you can do

  • List supported providers — GET /voices/providers
  • List a provider’s TTS models — GET /voices/{provider}/models
  • Search voices across all providers with filtering and pagination — GET /voices
  • Browse a single provider’s catalog — GET /voices/{provider}
  • Get a single voice’s metadata + preview MP3 — GET /voices/{provider}/{voiceId}
  • Buy / import / release Twilio phone numbers — POST /utils/buy-twilio-number, POST /utils/import-twilio-number, POST /utils/twilio/release-number
  • Repair Twilio webhooks — POST /utils/twilio/check-number
  • Assign a Twilio number to an agent for SMS — POST /utils/twilio/sync-sms

​
Supported providers

elevenlabs, deepgram, cartesia, rime-ai, openai, google-cloud, google-live (Gemini Live), ultravox, minimax, playht, azure. Every voice returned across providers is normalized to the same shape: 
{
  "voiceId": "21m00Tcm4TlvDq8ikWAM",
  "name": "Rachel",
  "provider": "elevenlabs",
  "previewUrl": "https://...mp3",
  "filters": {
    "accent": "American",
    "gender": "female",
    "useCase": "narration",
    "language": "en"
  }
}

​
Workspace API keys

Each provider can be used with either:
  1. The platform’s server-side API key (default fallback), or
  2. Your own workspace key — saved under workspaceData.secrets.<KEY_NAME> (e.g. ELEVENLABS_API_KEY, CARTESIA_API_KEY).
GET /voices/providers returns the exact workspaceSecretKey to use for each provider, plus a requiresWorkspaceApiKey flag for providers without a platform fallback. The voice listing endpoints (GET /voices and GET /voices/{provider}) all support:
ParamExampleBehavior
languageen / en-US / English / deInclusive match — en matches en-US, en-GB, English, etc.
gendermale / female / neutralCase-insensitive; aliases masculine / feminine / m / f accepted
accentamerican / britishCase-insensitive substring match
modelIdaura-2 / eleven_multilingual_v2Filters voice catalog for providers with model-specific voices
limit100 (default)1-500
offset0 (default)For pagination
providerselevenlabs,cartesia (unified)Comma-separated provider slugs (unified endpoint only)