dsh-voice-say@0.6.3
dsh-voice-say:把一句话合成语音。两条通道(本地 Breeze TTS 2 / 云端 apiKey+voiceId)共用 instructions + guidanceScale;18 个风格可叠加;合成前由「读音导演」重排标点——默认在 DeepSeek 窗口里的专用会话里想,按人设、当下情绪与对话情节断句;支持上传本地音色、自定义参考音频与试听片段。QQ 桥接会先问插件,群里发出去的语音与设置页试听同源同风格。 导演台连不上时自动改用「当前访问地址」自愈。
AI security review
scanned · by lpm-firewall-aiThe package’s default director is part of its speech-generation feature, and its README documents the API endpoint, key setting, and fallback when DSH is unavailable. The source sends the prompt and API key to that endpoint for the stated model request; no unrelated collection or attack behavior was identified.