Voice conversation mode for the DSH Web GUI: tap the mic to start a full-screen call, talk hands-free, and hear each reply read aloud as it streams. Bilingual (Chinese/English) UI, light and dark themes, adjustable speech rate, and switchable voices.

Install

# from npm (prebuilt)

dsh plugin --profile web add dsh-voice-talk

# from GitHub (first run asks for allowBuilds approval — follow the hint, retry)

dsh plugin --profile web add github:duoduoqian708/dsh-voice-talk

Any plugin you install runs third-party code with your own permissions — it can read your files, use your credentials, and reach the network, and tool approvals don’t sandbox it. GitHub-sourced plugins also run build scripts at install time — pnpm blocks those until you allow them, so an install can stop with ERR_PNPM_GIT_DEP_PREPARE_NOT_ALLOWED or ERR_PNPM_IGNORED_BUILDS; dsh prints the exact key to add under allowBuilds in your profile’s pnpm-workspace.yaml, and the install works on the next run. Allowing a build is a trust decision: only install sources you trust, and pin a commit (github:owner/repo#sha).

README

Voice chat mode for DSH Web: tap the mic to enter a full-screen call, speak and listen as you go, and replies are read aloud as they stream. Bilingual UI, light/dark theme aware, with speech-rate and voice controls.

English | 中文

Call mode

Quick start

1. Install

dsh plugin --profile web add dsh-voice-talk
dsh web   # restart after install

Install from source:

git clone https://github.com/duoduoqian708/dsh-voice-talk.git
cd dsh-voice-talk && npm install && npm run build
dsh plugin --profile web add /path/to/dsh-voice-talk

Requires dsh ≥ 0.1.0-rc.7 (web profile), Chrome / Edge, microphone permission and a network connection.

2. Configure a model

Get an API Key from Alibaba Cloud Model Studio, then paste it into Qwen's "Settings" under Settings → Plugins → Voice chat; the model and endpoint are built in. Qwen is the only engine for now; more will be added later.

3. Start talking

There is a mic icon next to the input box on the conversation page or workspace; tap it to enter the full-screen call and just start speaking. The red hang-up button or Esc exits; you can mute or collapse the panel mid-call.

Microphone entry next to the input box

Settings

Changes on the settings page are persistent defaults; changes made in the call overlay apply to the current call only.

Setting	Description	Default
Barge-in while speaking	Interrupt the readout by speaking; headphones recommended	Off
Auto-send after a pause (s)	How long a pause submits the utterance	2
Voice wave	Call-overlay waveform style: equalizer / ripple	Ripple
Rate	Readout speed, default per engine, adjustable per session	1.0x

Settings page

Highlights

Follows the system theme: light / dark out of the box.
Bilingual UI (Chinese / English): follows the platform language setting.
Keys stay on your machine and are never uploaded: the browser never sees them, and speech requests are proxied by the host.
Open source under MIT: auditable and self-hostable.

Easter egg

Suppose you would rather not stare at a screen and feel like getting outdoors; suppose you happen to have a proxy port open — then take your phone out for a walk and talk through your development while strolling.

Feedback

Found a problem or have an idea? Open an issue.

License

MIT

Content from the project README on GitHub ↗

Links

GitHub repository ↗ npm package ↗ Install via dsh-market ↗

More in this category

View the whole category →

Community comments

Comments are public GitHub Discussions. Loading them connects to GitHub and Giscus; a GitHub account is required to post.

duoduoqian708/dsh-voice-talk