DeepSeek Harness plugin

dsh-plugin-lmstudio

DSH (DeepSeek Harness) 插件:一键接入 LM Studio 本地推理服务器作为对话模型(OpenAI 兼容 /v1 端点)。安装后模型选择器出现 LM Studio 本地模型,不消耗云端 API 额度。

Jump to install

Source facts

Repository
starlightzfy/dsh-plugin-lmstudio
Latest update
Aug 16, 2026
Category
Models & Providers
GitHub stars
0
Format
plugin
Catalog evidence
Upstream dsh.bundle evidence
Evidence path
package.json#dsh.bundle
Checked against
0.1.0-rc.8
Upstream check date
2026-08-20

This evidence comes from the upstream catalog. This site has not installed, run, or security-reviewed the plugin.

Install

Start with a prompt that asks an agent to review the GitHub repository and source. Switch to the command if you want to install it yourself.

Copy this prompt into DSH, Codex, or another agent and ask it to review the GitHub repository and source first.

Do not install or run any commands yet. Read this plugin's GitHub repository, README, and relevant source code. Then answer the questions below clearly and directly so I can decide whether it fits my needs:

1. What is this plugin, and what problem does it solve?
2. Who is it for, and what are its typical use cases?
3. How is it used after installation? Include one minimal example.
4. What known limitations or privacy, security, compatibility, or maintenance risks does it have?
5. Give a clear recommendation: recommend, conditionally recommend, or do not recommend, with reasons.

Distinguish statements documented by the repository, inferences from source code, and unknowns. If evidence is insufficient, say so explicitly. Do not guess or simply repeat the README.

GitHub: https://github.com/starlightzfy/dsh-plugin-lmstudio
Plugin: dsh-plugin-lmstudio
Author: starlightzfy

Check the source files

Read the README and other files from this plugin directory before installing.

File explorer4 files
README.mdSource · read only
README language

dsh-plugin-lmstudio

Use a model served by LM Studio as a chat model inside the DeepSeek Harness (DSH) web GUI.

LM Studio exposes an OpenAI-compatible endpoint (http://127.0.0.1:1234/v1 by default). This plugin seeds DSH's built-in llm-pi-ai adapter with that route, so the model picker shows your local model — no cloud API quota consumed.

中文说明 · MIT

Requirements

  • DSH CLI + pnpm (see the DSH docs)
  • LM Studio with Developer → Start Server running (default port 1234)
  • At least one model loaded in LM Studio

Install

dsh plugin --profile web add dsh-plugin-lmstudio

The package declares dsh.bundle, so dsh plugin automatically adds it to the profile's bundle layers. Then restart the web app (dsh web) and refresh http://127.0.0.1:3080.

Configuration

The plugin ships a default provider route under the lm-studio provider key:

FieldDefaultMeaning
baseURLhttp://127.0.0.1:1234/v1LM Studio OpenAI-compatible endpoint
apiopenai-completionsWire protocol
modelsone example modelReplace with what you actually load

The bundled route is a base layer: anything you write under the llm-pi-ai: section of $DSH_HOME/settings.yaml merges over it per provider, so your own models and endpoints win. No API key is required for a local LM Studio server; if you enable key auth in LM Studio, add apiKeyEnv or an Authorization header in your settings layer.

Point it at your actual model

1. In LM Studio, load the model you want. 2. Find the exact model id: curl http://127.0.0.1:1234/v1/models and read the id field (usually owner/name, e.g. qwen/qwen3.5-9b). 3. Either edit the route in Settings → Models (the "ask the endpoint" button discovers models for you), or override it in $DSH_HOME/settings.yaml:

llm-pi-ai:
  providers:
    lm-studio:
      models:
        - id: qwen/qwen3.5-9b
          name: Qwen3.5 9B (LM Studio)
          contextWindow: 8192
          maxTokens: 4096

settings.yaml hot-reloads — no dsh web restart needed.

Usage

Refresh the page, open the model picker above the composer, and choose your LM Studio model (it appears under "LM Studio (本地)"). Traffic goes to 127.0.0.1:1234 only; your DeepSeek API balance is untouched.

> Tip: for plain conversation without any agent tools, create a "chat" agent > preset (an empty agent.cordis.yml) and pick it on the new-session screen — > that applies to every model, local or cloud.

License

MIT