Voice cloning

Voice cloning takes a short audio sample and returns a voice_id you can pass to /ai/text-to-speech for any future synthesis.

Capture a clean sample.

Record 10 seconds of a single speaker in a quiet environment. Use one of WAV, MP3, M4A, or FLAC at 16 kHz or higher. See Audio inputs for full guidance.
Send it to /ai/voice-clone.
- curl
- Python
Terminal window
curl -X POST https://api.gocommotion.com/ai/voice-clone \ -H "X-API-Key: eak_live_your_key_here" \ -F "file=@sample.wav" \ -F "name=narrator-en"
import requests with open("sample.wav", "rb") as audio: response = requests.post( "https://api.gocommotion.com/ai/voice-clone", headers={"X-API-Key": "eak_live_your_key_here"}, files={"file": audio}, data={"name": "narrator-en"}, ) voice_id = response.json()["voice_id"]
The response includes a stable voice_id like vc_01HXYZ....

Use the voice_id in synthesis.

curl -X POST https://api.gocommotion.com/ai/text-to-speech \
  -H "X-API-Key: eak_live_your_key_here" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Hello from your cloned voice.",
    "voice_id": "vc_01HXYZ..."
  }' \
  --output greeting.mp3