dsh-voice-input: A Voice Input Plugin for DeepSeek Harness Web UI

forrestahha/dsh-voice-input is a speech-to-text plugin for the DeepSeek Harness Web UI, released under the MIT license. It adds a microphone entry in front of the composer's send button and uses the browser's Web Speech API to stream recognition results into the current draft, without automatically sending the message—users confirm and send manually. The plugin supports both the standard and WebKit-prefixed SpeechRecognition interfaces, defaults to Chinese as the fallback language, and optimizes whitespace handling for both CJK and Latin text. Installation requires no additional API keys or backend services; the plugin can be added to a Web profile via the `dsh plugin` command. It is primarily intended for use with Chromium and other browsers that support