AgentEarth
登录

Ae · Qwen · 3 · Tts · Text · To · Speech · 1 · 7b

108积分/次
ae_qwen_3_tts_text_to_speech_1_7b

使用所选声音或提供的声音配置文件将文本转换为语音，并支持语言、风格和高级声音控制。

文本转语音音频
108积分/次

价格说明

108积分/次

API 调用

通过统一工具网关发起同步请求，鉴权和计费由 AgentEarth 处理。

cURL
curl --request POST \
  --url 'https://api.agentearth.ai/v1/tools/run' \
  --header 'Authorization: Bearer $AGENTEARTH_API_KEY' \
  --header 'Content-Type: application/json' \
  --data-raw '{
  "tool": "ae_qwen_3_tts_text_to_speech_1_7b",
  "input": {
    "text": "xxx",
    "top_k": "50",
    "top_p": "1",
    "voice": "xxx",
    "prompt": "xxx",
    "language": "Auto",
    "temperature": "0.9",
    "max_new_tokens": "200",
    "reference_text": "xxx",
    "subtalker_top_k": "50",
    "subtalker_top_p": "1",
    "repetition_penalty": "1.05",
    "subtalker_dosample": "true",
    "subtalker_temperature": "0.9",
    "speaker_voice_embedding_file_url": "xxx"
  }
}'
请求方式
POST
接口地址
/v1/tools/run
鉴权方式
Bearer Token
超时时间
60 秒
当前版本
—
响应示例
{
  "error_no": 0,
  "error_msg": "",
  "result": [
    {
      "type": "text",
      "text": "xxx"
    }
  ],
  "platform_annotations": {
    "source": "AgentEarth",
    "platform_note": "xxx",
    "platform_suggestion": "xxx"
  }
}

工具参数

未传入的可选参数将使用平台默认值。

参数类型必填说明默认值
textstring是The text to be converted to speech.—
top_kinteger否Top-k sampling parameter.50
top_pnumber否Top-p sampling parameter.1
voicestring否The voice to be used for speech synthesis, will be ignored if a speaker embedding is provided. Check out the **[documentation](https://github.com/QwenLM/Qwen3-TTS/tree/main?tab=readme-ov-file#custom-voice-generate)** for each voice's details and which language they primarily support.—
promptstring否Optional prompt to guide the style of the generated speech. This prompt will be ignored if a speaker embedding is provided.—
languageenum (Auto | English | Chinese | Spanish | French | German | Italian | Japanese | Korean | Portuguese | Russian)否The language of the voice.Auto
temperaturenumber否Sampling temperature; higher => more random.0.9
max_new_tokensinteger否Maximum number of new codec tokens to generate.200
reference_textstring否Optional reference text that was used when creating the speaker embedding. Providing this can improve synthesis quality when using a cloned voice.—
subtalker_top_kinteger否Top-k for sub-talker sampling.50
subtalker_top_pnumber否Top-p for sub-talker sampling.1
repetition_penaltynumber否Penalty to reduce repeated tokens/codes.1.05
subtalker_dosampleboolean否Sampling switch for the sub-talker.true
subtalker_temperaturenumber否Temperature for sub-talker sampling.0.9
speaker_voice_embedding_file_urlstring否URL to a speaker embedding file in safetensors format, from `fal-ai/qwen-3-tts/clone-voice` endpoint. If provided, the TTS model will use the cloned voice for synthesis instead of the predefined voices.—