Audio · OpenAI

Voice Creation

Create custom voices from audio samples for use with OpenAI TTS models.

POST/v1/audio/voices/create

Custom voices work with the gpt-4o-mini-tts model. Pass the returned voice ID as { "id": "voice_abc123" } in the TTS voice parameter.

Request

Multipart Form Parameters

audio_samplefileformrequired

Audio file with a sample of the voice to clone

consentstringformrequired

Consent confirmation text

namestringformrequired

Name for the custom voice

descriptionstringform

Optional description of the voice

cURL

curl https://api.metriqual.com/v1/audio/voices/create \
  -H "Authorization: Bearer mql_your_key" \
  -F audio_sample=@my_voice.mp3 \
  -F consent="I confirm I have the right to use this voice" \
  -F name="My Custom Voice"

TypeScript SDK

const voice = await mql.audio.createVoice({
  audio_sample: audioFile,
  consent: 'I confirm I have the right to use this voice',
  name: 'My Custom Voice',
});

console.log('Voice ID:', voice.id);

// Use in TTS
const audio = await mql.audio.speech({
  model: 'gpt-4o-mini-tts',
  input: 'Hello from my custom voice!',
  voice: { id: voice.id },
  instructions: 'Speak in a friendly, natural tone.'
});

Python SDK

with open("my_voice.mp3", "rb") as f:
    voice = mql.audio.create_voice(file=f, name="My Custom Voice")
print("Voice ID:", voice["id"])

# Use in TTS
audio = mql.audio.speech(
    input="Hello from my custom voice!",
    voice=voice["id"],
    model="gpt-4o-mini-tts",
)

Response

200

{
  "id": "voice_abc123def456",
  "created_at": 1719000000,
  "name": "My Custom Voice",
  "object": "audio.voice"
}

Quick Start