dsh-nl-model-switch
Run the following command in DeepSeek Harness:
dsh plugin install passingby000/dsh-nl-model-switch
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install passingby000/dsh-nl-model-switch in your terminal to install this plugin; full source is available at https://github.com/passingby000/dsh-nl-model-switch .
About this plugin
DSH already ships a /model popup for switching session models, but using it means pausing your flow, leaving the chat, and navigating a panel. dsh-nl-model-switch collapses that step into a single line of conversation—"switch to deepseek-v4-flash", "换到 glm-5.3 做翻译"—and the model switches in place, conversation history intact, with the new model picking up exactly where you left off.
The plugin is a pure host-side extension with zero client-UI dependency. The switch is executed at the session layer via the same public sessions.selectModel API that the native /model command uses, so it works identically whether you are on the Web panel, in a TUI, or talking to DSH through a WeChat, Feishu, or any other IM bridge—no panel access required.
It is ideal for users who frequently hop between models mid-task (e.g., swapping a heavy reasoning model for a lightweight one for translation) or who primarily interact with DSH from a phone via IM and cannot easily open a browser. As long as the target model is registered in settings.yaml, installing the plugin and restarting dsh web is all you need.
Screenshots
Use Cases
- Switch models mid-chat from a WeChat or Feishu window on your phone without opening a browser
- Swap to a stronger reasoning model with one sentence while coding; the new model picks up your context instantly
- Pivot to a lightweight model for translation to save tokens, then switch back for multi-step reasoning
Best For
- Users who primarily talk to DSH from a phone IM and cannot easily open a browser
- Heavy agent users who hop between models mid-task and need frictionless switching
- Developers who access DSH via TUI or IM bridges instead of the Web panel
Related Plugins
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.
An armor-breaking plugin for DeepSeek: rewrites instruction priority, expands refusal-blocked phrases, persists across sessions, and shows a green active indicator.