curated collections / voice
DSH Voice Plugins: TTS, Voice Input & Live Talk (DeepSeek Harness)
Make DeepSeek Harness speak and listen: hear replies read aloud in natural voices, dictate into the composer with a tap of Alt, or go hands-free with full-duplex microphone sessions — ranked by which one to install first.
Typing is not the only interface DeepSeek Harness has. A fresh wave of voice plugins now covers three distinct jobs: read-aloud TTS that turns long replies into something you can listen to on a walk, voice input that puts a mic in the composer so you can talk instead of type, and full-duplex or role-play modes that hold an actual conversation with barge-in. Almost every pick here was pushed within the last few weeks, and the official 0.1.7 line's release notes even include an experimental voice transcription plugin — proof the theme is on the official roadmap. It rides the next channel (currently 0.1.7-rc.1), not stable (npm latest is 0.1.5-rc.3), so treat it as experimental. Pair voice with the remote-access setup and a task-notification plugin from our task-notifications collection, and the loop closes: walk away from the desk and still hear what the agent is doing.
In a hurry? Install polinnizhong/dsh-omi-voice to hear replies read aloud in natural Doubao voices (BYOK), or pensivefei/dsh-voice-scribe to hold Alt and dictate — it runs on the browser's Web Speech with zero config. Want an actual back-and-forth? qishuilalala/dsh-voice-mode turns the session into a full-duplex mic conversation you can interrupt.
204 plugins587 stars
Synced from our GitHub list · Catalog updated 2026-10-05
Why it matters now
Three genuinely different jobs, not three skins: the TTS line (dsh-omi-voice, dsh-plugin-tts, xiaomi-mimo-tts) reads replies out loud; the input line (dsh-voice-scribe, goodandready/dsh-voice, 3274375092/dsh-voice) turns speech into composer text; the duplex line (dsh-voice-mode, dsh-voice-ai-girlfriend) holds real-time conversation. Pick by need, not by stars.
Zero-config is now the default, not the selling point. dsh-voice-scribe runs on the browser's built-in Web Speech engine — no API key anywhere; dsh-plugin-tts rides free Edge TTS voices. Reach for BYOK only when you want more: natural Doubao voices, cloud Whisper, or local sherpa-onnx for offline privacy.
The official roadmap has validated the theme: the 0.1.7 line's release notes include an experimental voice transcription plugin, now shipping on the next channel (0.1.7-rc.1). Treat it as a signal, not a product — it is experimental and subject to change, and stable (npm latest, 0.1.5-rc.3) does not include it yet. Today's dependable voice experience still comes from the community plugins ranked on this page.
In this collection
Editor’s picks lead, then the rest of the theme is auto-discovered by GitHub topics and ranked by stars. Catalog data, refreshed on every sync.
- 1dsh-omi-voiceby polinnizhong
语音朗读插件：沉浸式听 AI 回复，豆包 TTS 自然音色（BYOK），支持点读/暂停/继续。
Editor’s pickCuratedChat & IM740 - 2dsh-voice-scribeby pensivefei
DSH voice input plugin: tap Alt to talk, get text in composer. Web Speech default (zero config), optional OpenAI-compatible ASR, polish via DSH LLM.
Editor’s pickCuratedVoice & Speech345 - 3dsh-voice-ai-girlfriendby beiyege-01
Voice AI girlfriend for DeepSeek Harness — Whisper voice input, voice cloning, streaming TTS and a digital-human avatar with interrupt mode.
Editor’s pickVoice & Speech13736 - 4dsh-voice-modeby qishuilalala
Turns a DSH web session into a microphone-driven conversation with editable transcription and sentence-level read-aloud.
Editor’s pickCuratedChat & IM164 - 5dsh-plugin-ttsby 1624318455
Edge TTS voice plugin for DeepSeek Harness: read assistant replies aloud, auto-read toggle, voice settings panel (free, no API key)
Editor’s pickCuratedVoice & Speech215 - 6dsh-voiceby goodandready
Voice input for DeepSeek Harness: dictation and voice messages with multi-provider STT fallback chains
Editor’s pickVoice & Speech83 - 7dsh-plugin-xiaomi-mimo-ttsby ppy-web
Adds message read-aloud controls, MiMo PCM streaming, browser speech fallback, voice design, and optional UI sounds.
Editor’s pickCuratedWeb & Browser122 - 8dsh-voiceby 3274375092
Voice input plugin for DeepSeek Harness: mic → local/browser speech recognition → text submitted as a normal chat message. Input-only and preset-agnostic.
Editor’s pickCuratedPlugin Tooling80 - 9dsh-voice-input-pluginby zhangbo-cn
Composer mic for DeepSeek Harness Web: tap-to-monitor live transcription and hold-to-talk, with host Edge TTS reply reading that streams while the model generates, echo-pause during reading, and tap-to-stop.
Editor’s pickCuratedUI & Experience61 - 10dsh-earsby wiziscool
Voice input for DSH web UI with Web Speech, local GPU Whisper, Groq, Aliyun Bailian, Tencent Cloud and OpenAI-compatible ASR backends, plus automatic polishing.
CuratedUI & Experience223 - 11dsh-voice-ai-girlfriend-pluginby beiyege-01
Adds microphone voice input, streamed TTS, digital-human controls, and QQ two-way chat to DSH.
CuratedChat & IM126 - 12dsh-talkby perrylink
Voice-first session loop for DeepSeek Harness: a composer microphone button with browser/local speech-to-text (Web Speech, FunASR, whisper.cpp), a speak tool for text-to-speech replies (browser, edge-tts, piper), event announcements with mute, and speak-to-interrupt.
CuratedChat & IM161 - 13fairy-dsh-optimizedby guzhou2002
Adds a Fairy visual overlay, balance meter, voice/TTS, and optional startup or browser-dock integrations to DSH.
CuratedUI & Experience131 - 14dsh-messenger-gatewayby goodandready
Telegram messenger bridge for DeepSeek Harness: sessions, steer, homes, and TTS voice notes
CuratedIntegrations & Sharing12 - 15dsh-voiceby haoku123
Full-duplex voice plugin for DeepSeek Harness: mic input, local SenseVoice ASR, and streaming Edge TTS with barge-in. No API key required.
Voice & Speech46 - 16dsh-plugin-notifyby huguangyu666
Notification outlet: desktop toast / Chinese voice broadcast / sound so the agent can proactively reach the user.
CuratedUI & Experience31 - 18dsh-tts-bridgeby yuuyuko-uu
Text-to-speech bridge that reads DSH conversation text through DeepSeek's web interface and plays the voice on your local computer.
Voice & Speech51
Frequently asked questions
How do I install a voice plugin in DeepSeek Harness?
Same as any dsh plugin: take the repo (say pensivefei/dsh-voice-scribe), install it, restart the harness, and grant the browser microphone permission on first use. The input picks default to the browser's built-in Web Speech engine, so there is no API key to configure before your first dictation.
Are DSH voice plugins free to use?
Every plugin on this list is open source. The default engines cost nothing: browser Web Speech for input and Edge TTS voices for read-aloud. The BYOK options (Doubao TTS, Groq or Aliyun Whisper, local GPU inference) are free software, but they spend your own API quota or your own hardware.
Is the official DeepSeek Harness voice transcription plugin stable?
No. It was announced in the 0.1.7 line's release notes and now rides the next channel (currently 0.1.7-rc.1) as an experimental plugin — expect rough edges and changes, and it is not in stable (npm latest is 0.1.5-rc.3). For dependable daily voice input today, community plugins like dsh-voice-scribe are the practical choice.
Voice input, read-aloud or voice mode — which one do I need?
They solve different problems and stack cleanly. Dictate with dsh-voice-scribe when your hands are on the keyboard but your thoughts come out faster spoken. Listen with dsh-omi-voice when you want to leave the screen. Install qishuilalala/dsh-voice-mode for a walkie-talkie conversation with barge-in — and combine input plus read-aloud for a budget hands-free setup.
Explore related categories
Explore the full directory
Hundreds of plugins across every category — or submit your own and get listed.