Add Vocals music layering API

提供纯音乐录音

选择来源编配并准备可访问 upload_url，保留原音频，使新带人声版本能够结合目标音乐上下文进行检查。

{
  "model": "add-vocals",
  "input": {
    "upload_url": "https://cdn.doculator.org/audio/FY414XS6JCUXLDGW/FY414XS6JCUXLDGW_1.mp3",
    "prompt": "[Verse]\nRain has left the window bright\nMorning folds away the night\nOn the table, bread and tea\nThere is room for you and me\n[Chorus]\nOpen up the day again\nLet the sunlight wander in\nEvery road can start from here\nOne small step and skies are clear",
    "title": "Harbor Morning Song",
    "style": "Warm acoustic folk, clear intimate lead vocal, gentle hopeful phrasing",
    "negative_tags": "harsh distortion, shouted vocals",
    "mv": "V5"
  }
}

准备歌词或人声方向

在 prompt 中编写目标歌曲文字或完整人声要求，再根据这些指令检查完成人声乐句与歌词处理。

描述音乐演唱风格

通过 style 表达曲风与演唱方向，为结果设置标题，并通过必填 negative_tags 描述希望弱化的元素。

{
  "model": "add-vocals",
  "input": {
    "upload_url": "https://cdn.doculator.org/audio/FY414XS6JCUXLDGW/FY414XS6JCUXLDGW_2.mp3",
    "prompt": "[Verse]\nLanterns trace the evening tide\nSmall boats turn toward the light\n[Chorus]\nCarry home the quiet sea\nLeave a little room for me",
    "title": "Lanterns on the Tide",
    "style": "Reflective acoustic folk, soft lead vocal, quiet coastal evening mood",
    "negative_tags": "harsh distortion, shouted vocals",
    "mv": "V5"
  }
}

使用接受的偏好字段

需要时准备适用 vocal_gender 偏好与支持的音乐权重，并在检查结果时将这些设置与完整提示词和来源一并保存。

检查完成的带人声音乐

读取返回候选并结合来源纯音乐聆听，再保存选定音频地址与标识，用于另一项编辑、续接或项目演示。