make-dsh-voice
Run the following command in DeepSeek Harness:
dsh plugin install blueperformer/make-dsh-voice
Paste the following prompt into your AI chat to install this plugin:
Install in DeepSeek Harness by running dsh plugin install blueperformer/make-dsh-voice; source code is available at https://github.com/blueperformer/make-dsh-voice .
About this plugin
In the DSH Web UI every agent reply is a wall of plain text. In a long conversation your eyes are pinned to the screen, and there is no natural listen path.
make-dsh-voice closes that gap. It calls the Alibaba Cloud DashScope cosyvoice model to synthesize each agent reply into a short audio clip in a voice you have cloned, then docks a single-line player right under the composer. A settings page (Settings > Voice) lets you switch the API key, synthesis model, enrolled voice ID, reply length policy (short / auto / full), and output folder; all values hot-reload so the very next turn picks up the change. A toggleable boot sound plays on the first user gesture after the page loads.
Best for Windows 11 users running the DSH web profile who already have a DashScope API key and an enrolled voice ID and would rather listen through long replies than read them word by word, as well as for plugin authors looking for a real-world reference implementation of a voice-synthesis outlet inside the DSH plugin system.
Use Cases
- Listen through long agent replies instead of reading them word by word
- Dock a single-line voice player under the DSH composer that follows each newest clip
- Have the agent speak its replies in your own cloned voice via DashScope
Best For
- Windows 11 users running the dsh web profile who want a listen-through path for replies
- Users with a DashScope API key and enrolled voice ID who prefer spoken answers
- Plugin authors building voice-synthesis outlets for DSH who want a working reference implementation
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.