v0.1.0

First public release of dsh-ears.

Voice input for DeepSeek Harness:

microphone → transcription → optional polish → editable draft → manual send

Install

dsh plugin --profile web add dsh-ears

Without the dsh CLI:

npx -y @deepseek-ai/dsh plugin --profile web add dsh-ears

Compatibility

• DeepSeek Harness 0.1.0-rc.6 or 0.1.0-rc.7
• Node.js ^22.19.0 || >=24.0.0

Surface

• Composer microphone plus a recognition bar with waveform, stop, and discard
• Backends: browser Web Speech, local Whisper, Groq, Alibaba Cloud Model Studio, custom OpenAI-compatible
• Optional polish through any model already configured in dsh
• Default shortcut Ctrl+Shift+Space
• Native settings page (General, Recognition, Polishing, About)

See CHANGELOG.md for the full first-release notes.

npm: https://www.npmjs.com/package/dsh-ears