AI Agent Hub
Back to plugins
dsh-speech preview

dsh-speech

Client Updated 2026.09.04

Run the following command in DeepSeek Harness:

dsh plugin install allmodels-io/dsh-speech

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install allmodels-io/dsh-speech in your DeepSeek Harness terminal, then configure your AllModels.io account under Settings → Speech after restart, source: https://github.com/allmodels-io/dsh-speech

About this plugin

Hours of typing instructions and scrolling through output are the two biggest frictions in any coding-agent session. dsh-speech brings the most natural interface back: speak your requests aloud, then listen to a concise spoken recap once the work is finished, so your hands and eyes can rest from the screen.

Two voice pipelines form the core. On input, tap the microphone next to the send bar and start talking; the transcript appears in real time, ready to edit before you hit send. On output, when the agent delivers its final message the plugin distills it into a short spoken summary. Enable autoplay to hear it the moment it is ready, or press play later at your convenience. All transcription and speech synthesis is powered by AllModels.io, whose starter credit covers several hours of usage.

It is built for developers who live inside DeepSeek Harness and want to cut down keystroke and screen-reading fatigue, or simply prefer to confirm task completion by ear while their attention is elsewhere.

Screenshots

Use Cases

  • Dictate instructions to your coding agent without typing
  • Listen to a spoken summary while your eyes are off-screen
  • Transcribe speech, edit quickly, then send the refined prompt

Best For

  • Developers who live inside DeepSeek Harness daily
  • Coders who want to cut down keyboard and screen fatigue
  • Users who prefer voice-driven workflows