Skip to main content
Method: client.generateSpeech().

​
Example

const audio = await client.generateSpeech({
  model: "openai/gpt-4o-mini-tts",
  input: "Hello world",
  voice: "alloy",
  response_format: "mp3",
});

​
Key parameters

  • model (required): TTS-capable model id.
  • input (required): Text to synthesize.
  • voice: Provider-normalized voice alias (recommended) or provider-native voice value.
  • response_format: mp3, wav, flac, etc.
  • speed: Playback speed multiplier (typically 0.25-4).
If voice is not valid for the routed provider/model mapping, the API returns 400 with param: "voice".

​
Provider config examples

// ElevenLabs explicit override
await client.generateSpeech({
  model: "eleven-labs/eleven-v3",
  input: "Hello from ElevenLabs",
  voice: "rachel",
  config: {
    elevenlabs: {
      voice_id: "21m00Tcm4TlvDq8ikWAM"
    }
  }
});

// Google explicit override
await client.generateSpeech({
  model: "google/gemini-2.5-flash-preview-tts",
  input: "Hello from Gemini",
  voice: "kore",
  config: {
    google: {
      voice_name: "Kore"
    }
  }
});

​
Returns

Audio binary (Blob)
Last modified on February 25, 2026