stardustlc666/dsh-voice
Voice two-pack: voice_tts (edge-tts protocol free Microsoft Neural Voice, Sec-MS-GEC native DRM generation) / voice_stt (OpenAI compatible with ASR) / voice_list, WS compression + optional proxy tunneling
dsh-voice is a voice two-pack for DeepSeek Harness: voice_tts converts text to speech via the edge-tts protocol (Microsoft Edge read-aloud service — free, unlimited, 22+ built-in voices), and voice_stt converts speech to text through any OpenAI-compatible ASR endpoint (Groq / OpenAI / custom), with voice_list exposing the available voice catalog. voice_tts works with zero configuration; voice_stt needs an ASR API key configured via the plugin settings (asrEngine, asrModel, asrApiKey).
Install
dsh plugin --profile web add dsh-voicenpm dsh-voice 0.3.1 verified 2026-09-01 (maintainer stardustlc; README has EN edition, repo stardustlc666/dsh-voice). Voice two-pack verified on @deepseek-ai/dsh@0.1.2-alpha.2 (2026-08-31); no runtime imports of @deepseek-ai/* internals. voice_tts is zero-config; voice_stt needs an ASR key. Restart the web service after install.
Compatibility
DSH @deepseek-ai/dsh 0.1.2-alpha.2 (verified); edge-tts protocol (free Microsoft Neural Voice, 22+ voices); OpenAI-compatible ASR (Groq/OpenAI/custom).
Details
- Repo: stardustlc666/dsh-voice
- Category: Infrastructure & Deployment
- Stars: 2
- Version: npm dsh-voice 0.3.1
- Last push: 2026-09-18
- First seen: 2026-08-14
Recent updates
voice_tts (edge-tts, free unlimited); voice_stt (OpenAI-compatible ASR); voice_list catalog; zero-config TTS.
FAQ
- Is text-to-speech free?
- Yes — voice_tts uses the edge-tts protocol (Microsoft Edge read-aloud), which is free and unlimited with 22+ voices.
- What do I need for speech-to-text?
- An ASR key — voice_stt uses OpenAI-compatible endpoints (Groq / OpenAI / custom) with a configurable model.
- How do I install it?
- dsh plugin --profile web add dsh-voice (npm 0.3.1) and restart the web service.
Alternatives
stardustlc666/dsh-sql · tangbut1/dsh-plugin-voice-input · zhangbo-cn-dsh-client-ui-voice-input