Speech›Fast Voice Clone
API Key not detected. Please configure to use debug features.
1Clone Audio (file_id)*

Supported formats: Mono WAV

Size limit: ≤ 20MB

Duration: 5s - 2mins; Recommended: 5-15s

2Configure Voice ID (voice_id)*
8-64 chars, start with a letter, use letters/numbers/-/_, cannot end with - or _. voice_id must be unique
3Prompt Audio (Optional)
4Audition Config (Optional)
Task Info
File ID-
Trace ID-
Status-
Guide
Recommended uploaded audio duration is at least 5 seconds and no more than 2 minutes.
Use clear, noise-free human speech recordings for both reference audio and style audio.
If style audio and its corresponding transcript are provided during cloning, the system can extract style and emotion features from the style audio and synthesize them with timbre information from the reference audio. The style audio and reference audio can come from different speakers.
Speech synthesis can use the u2-tts-clone model, and you can call the cloned voice with the corresponding voice_id.
API Reference
CURL
curl -X POST "https://maas-api.hivoice.cn/v1/audio/voices/clone" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "file_id": ,
    "voice_id": "",
    "model": "u2-tts-clone",
    "clone_prompt": {
      "prompt_audio": ,
      "prompt_text": ""
    },
    "text": ""
  }'
Request Body JSON
{
    "file_id": ,
    "voice_id": "",
    "model": "u2-tts-clone",
    "clone_prompt": {
      "prompt_audio": ,
      "prompt_text": ""
    },
    "text": ""
  }