OpenSpeechAPI Control Center

Dashboard

syncing

Providers

-

Healthy

-

Running Tasks

-

Last Update

-

Capability Lab

STT - Speech to Text

Provider Mode Language

Advanced Options

Model Device Beam Size

Audio File

Result

TTS - Text to Speech

Provider Voice Speed

Advanced Options

Model Transport Language

参考音频 / Reference clip (声音克隆引擎必需,如 coqui XTTS) 参考文本 / Reference transcript (可选;填=上面音频说的原话→zero-shot 更准,留空=免转写跨语种克隆) Text

自动保存音频到历史

Result

Interaction Log

Verbose

Realtime 语音对话

idle

端到端实时语音对话(OpenAI gpt-realtime / Qwen-Omni realtime)。点击「开始对话」授权麦克风后直接与 AI 语音交谈,服务端 VAD 自动断句、AI 语音实时播放。需先在 Config 页配置 openai-realtime / qwen-omni-realtime 引擎。

Provider Model Voice

高级选项 / Advanced

断句模式 / Turn detection 输入转写 / Input transcription

System instructions (可选,给 AI 的系统提示)

对话将显示在这里 —— 你说的话与 AI 的回复。

会话状态 / Session

Engine Catalog

Click engine name to configure. Install/Uninstall toggles providers.yaml entry.

全局设置 / Global

HF Endpoint (HF_ENDPOINT) 默认下载源 / Default Model Source 模型存储根目录 / Model Storage Root (OSA_MODEL_ROOT)

Provider Configuration

Select a vendor or engine from the left to edit its settings.

Engine Task History

Persistent Docker/Native engine install, update, start, and stop tasks.

Time	Engine	Action	Status	Message

Control Center

Dashboard

-

-

-

-

Capability Lab

Result

Result

历史音频 (0)

Interaction Log

Realtime 语音对话

会话状态 / Session

Engine Catalog

全局设置 / Global

Provider Configuration

Engine Task History

Add Vendor