声音复刻

POST

/tk/v2/api/v1/services/audio/tts/customization

提交声音复刻任务（音色克隆训练）。

支持模型：

voice-enrollment（配合 cosyvoice-v3.5-plus / cosyvoice-v3.5-flash）

说明： 根据上传的音频样本训练目标音色，返回训练后的 voice_id。
qwen-voice-enrollment 已下线，请改用 voice-enrollment：action=create_voice，target_model 填写已开通的 cosyvoice-v3.5-plus 或 cosyvoice-v3.5-flash。创建成功后拿 voice_id，到「语音合成（标准 TTS）」合成，且合成时的 model 须与 target_model 一致。
MiniMax复刻见「语音合成(多模态TTS)」

请求参数

Header 参数

Body 参数application/json必填

示例

默认示例voice-enrollment 示例

{
    "model": "voice-enrollment",
    "input": {
        "action": "create_voice",
        "target_model": "cosyvoice-v3.5-plus",
        "prefix": "demo",
        "url": "https://dashscope.oss-cn-beijing.aliyuncs.com/samples/audio/cosyvoice/cosyvoice-zeroshot-sample.wav"
    }
}

请求示例代码

Shell

JavaScript

Java

Swift

PHP

Python

HTTP

Objective-C

Ruby

OCaml

Dart

cURL

默认示例voice-enrollment 示例

curl --location '/tk/v2/api/v1/services/audio/tts/customization' \
--header 'Authorization: Bearer {{api_key}}' \
--header 'Content-Type: application/json' \
--data '{
    "model": "voice-enrollment",
    "input": {
        "action": "create_voice",
        "target_model": "cosyvoice-v3.5-plus",
        "prefix": "demo",
        "url": "https://dashscope.oss-cn-beijing.aliyuncs.com/samples/audio/cosyvoice/cosyvoice-zeroshot-sample.wav"
    }
}'

返回响应

🟢200

application/json

成功

Bodyapplication/json

示例

{
    "output": {
        "voice_id": "demo_voice",
        "status": "SUCCESS",
        "preview_audio_url": "https://example.com/preview.wav"
    },
    "usage": {
        "characters": 0
    },
    "request_id": "94d9bda2-5c77-0f84-8d92-1c6c5f4e3b7a"
}

语音合成（多模态TTS）

原生Gemini格式