AITool Local Audio Gateway

快速测试

这里直接调用现有接口。测试失败时会显示 HTTP 状态和错误详情，方便判断是网关、鉴权还是上游模型服务的问题。

API Key（可选）

文字转语音测试

测试文本

模型音色格式

下载生成音频

语音转文字测试

模型语言（可选）

提示词（可选）音频文件

接口状态测试

音色列表测试

按模型过滤（可选）

功能和调用链接

Method	Path	用途	鉴权
GET	`/`	Open this dashboard page.	No
GET	`/health`	Health check.	No
GET	`/dashboard.json`	Machine-readable dashboard metadata.	No
GET	`/v1/models`	List configured models/backends.	Yes if API keys are configured
GET	`/v1/voices`	List voices, optionally filtered by ?model=chattts.	Yes if API keys are configured
POST	`/v1/audio/speech`	Text to speech. OpenAI-style JSON body.	Yes if API keys are configured
POST	`/v1/audio/transcriptions`	Speech to text. OpenAI-style multipart upload.	Yes if API keys are configured

Swagger 文档：http://127.0.0.1:8010/docs。外部调用时端口取决于你的内网穿透映射。

后端模型状态

Backend	Status	Upstream	Model Path	Aliases
Qwen3-TTS-12Hz-1.7B-CustomVoice `qwen_tts`	needs_base_url openai_speech_proxy	`Not configured`	`C:\01.Jiaobinshi\05.AI\Models\Qwen3\Qwen3-TTS-12Hz-1.7B-CustomVoice`	qwen-tts, qwen3-tts-1.7b-customvoice
Qwen3-ASR-1.7B `qwen_asr`	enabled openai_transcription_proxy	`http://127.0.0.1:8020/v1/audio/transcriptions`	`C:\01.Jiaobinshi\05.AI\Models\Qwen3\Qwen3-ASR-1.7B`	qwen-asr, qwen3-asr-1.7b
GPT-SoVITS v2Pro `gpt_sovits`	enabled gpt_sovits	`http://127.0.0.1:9880/tts`	`C:\01.Jiaobinshi\05.AI\Project\GPT-SoVITS-v2pro-20250604`	gpt-sovits, sovits
ChatTTS UI `chattts`	enabled chattts	`http://127.0.0.1:9966/tts`	`C:\01.Jiaobinshi\05.AI\Project\win-ChatTTS-ui-v1.0`	chat-tts
faster-whisper `faster_whisper`	disabled faster_whisper	`Not configured`	`large-v3`	whisper, faster-whisper

Qwen TTS/ASR 当前是 sidecar 模式；如果显示 needs_base_url，需要先启动兼容服务并在 config/gateway.json 填入 base_url。

可用音色

Backend	Voices
`qwen_tts`	AidenDylanEricOno_annaRyanSerenaSoheeUncle_fuVivian
`gpt_sovits`	reference
`chattts`	22227869665340995099

调用示例

PowerShell TTS

Invoke-WebRequest -Uri http://127.0.0.1:8010/v1/audio/speech -Method POST -ContentType "application/json" -Body '{"model":"chattts","input":"你好，这是本地语音网关测试。","voice":"2222","response_format":"wav"}' -OutFile speech.wav

curl TTS

curl.exe -X POST http://127.0.0.1:8010/v1/audio/speech -H "Content-Type: application/json" -d "{\"model\":\"chattts\",\"input\":\"你好，这是本地语音网关测试。\",\"voice\":\"2222\",\"response_format\":\"wav\"}" -o speech.wav

JSON body

{
  "model": "chattts",
  "input": "你好，这是本地语音网关测试。",
  "voice": "2222",
  "response_format": "mp3"
}

curl STT

curl.exe -X POST http://127.0.0.1:8010/v1/audio/transcriptions -F "model=qwen_asr" -F "file=@C:\path\audio.wav" -F "response_format=json"

启动命令

网关服务	`scripts\run_gateway.cmd`
GPT-SoVITS API	`scripts\start_gpt_sovits_api.ps1`
ChatTTS	`scripts\start_chattts.ps1`
Qwen-ASR	`scripts\start_qwen_asr_api.ps1`
资源诊断	`scripts\diagnose_resources.ps1`

本机资源路径

Name	Path
models_root	`C:\01.Jiaobinshi\05.AI\Models`
gguf_root	`C:\01.Jiaobinshi\05.AI\Models_GGUF`
gpt_sovits_root	`C:\01.Jiaobinshi\05.AI\Project\GPT-SoVITS-v2pro-20250604`
chattts_root	`C:\01.Jiaobinshi\05.AI\Project\win-ChatTTS-ui-v1.0`
comfyui_aki_root	`C:\01.Jiaobinshi\05.AI\Project\ComfyUI-aki-v3`
sd_webui_aki_root	`C:\01.Jiaobinshi\05.AI\Project\sd-webui-aki\sd-webui-aki-v4.10`