dsh-voice-input-plugin

In chat applications, typing can be inefficient or cumbersome, especially in multilingual environments or on mobile devices. This plugin brings voice control to DeepSeek Harness, allowing users to convert speech to text in real-time through simple gestures, boosting input speed and interaction quality.

The plugin’s core capabilities feature two intuitive voice input modes: tap-to-monitor enables continuous listening with live text streaming into the draft, letting users send anytime or keep adding speech; hold-to-talk offers precise control where releasing sends the message and automatically reads replies aloud. Technically, it runs entirely in the browser using the Web Speech API for recognition—no API keys needed—while leveraging Edge TTS for high-quality reply reading, ensuring accessibility.

This plugin is ideal for a broad range of DeepSeek Harness users, from everyday chatters to multilingual communicators and developers. It excels in scenarios where typing is inconvenient or efficiency is key, such as meeting notes, quick responses, or accessibility needs, providing seamless voice interaction without complex backend setups.