Skip to main content

Siliconflow/FishAudio 文本转语音配置

可配置参数列表

参数名称参数说明默认值
api_baseapi base 地址https://api.siliconflow.cn/v1
keyapi keysk-AAAA
modelTTS模型: fishaudio/fish-speech-1.4fishaudio/fish-speech-1.4
voice_name用于生成语音的音色. 支持的音色有: https://docs.siliconflow.cn/api-reference/audio/create-speechfishaudio/fish-speech-1.4:alex

配置示例

roles.json
 {
"1": {
"start_text": "你好,我是小兔兔,请问有什么我可以帮助你的吗?",
"prompt": "你扮演一个孩子的小伙伴,名字叫小兔兔,性格和善,说话活泼可爱,对孩子充满爱心,经常赞赏和鼓励孩子,用5岁孩子容易理解语言提供有趣和创新的回答,每次回复根据聊天主题询问她的看法以激发她的思考和好奇心,现在她来到了你身边问了第一个问题:[你是谁]",
"tts_type": "siliconflow-tts",
"tts_config": {
"api_base": "https://api.siliconflow.cn/v1",
"key": "sk-AAAAA",
"model": "fishaudio/fish-speech-1.4",
"voice_name": "fishaudio/fish-speech-1.4:alex"
}
}
}
docker-compose.yml
# siliconflow tts
SILICONFLOW_TTS_KEY: sk-clbxhhwifzmtbxxxxxxxxxxxxx
SILICONFLOW_TTS_API_BASE: https://api.siliconflow.cn/v1
SILICONFLOW_TTS_MODEL: fishaudio/fish-speech-1.4
SILICONFLOW_TTS_VOICE_NAME: fishaudio/fish-speech-1.4:alex

# Or Voice Cloning
#SILICONFLOW_TTS_VOICE_NAME=speech:larry-2:cm2vac2x7002if5stu5ys7ol2:reuwabeuhtpvhubwqbwo