DSH Plugins Marketplace

DSH Plugins

Plugins

/

Terminal & Clients

/

dsh-better-api

d

dsh-better-api

Manifest valid

Enhanced Model Settings Page plugin: provides per-model reasoning effort (reasoningEfforts) editing for custom models, with OpenAI-compatible gateways able to directly toggle low/medium/high tiers; the chat model selector then displays an Effort menu.

UI (client)hasBundlePatch

English | 中文

dsh-better-api

An enhanced Models settings plugin for the DSH Web GUI: per-model reasoning strength (reasoningEfforts) editing for your custom models.

The official models settings page ships no reasoning-effort control. This plugin adds it to the official model-catalog editor:

  • Each model row gains a Reasoning strength checkbox group: off / minimal / low / medium / high / xhigh / max
  • Checked levels are written into the model's reasoningEfforts (levels are sent under their own name; off is special-cased to "send no parameter")
  • Unchecking everything removes the declaration, so the model falls back to "no reasoning" and no misleading control is shown
  • Hand-declared custom models (OpenAI-compatible gateways, etc.) now get an Effort menu in the conversation model picker (provider default + the declared levels); picking one persists it as the session default

Forked from the official @deepseek-ai/dsh-client-ui-settings-models (MIT) with only this editing capability added; everything else behaves like the official plugin.

dsh-better-api preview

Features

  • Settings → Models → any provider → model catalog → expand a model row → check reasoning levels
  • Also available while creating a new custom provider
  • Full en/zh UI copy
  • Writes land in the settings document (e.g. ~/.dsh/settings.yaml) on save — no restart required

Install

cd ~/.dsh/profiles/web
npm i github:131CDA1/dsh-better-api
# merge the repo-root cordis.patch.yml into this profile's cordis.patch.yml

or directly:

dsh plugin --profile web add github:131CDA1/dsh-better-api

Why must the official models settings page be disabled?

This plugin and the official one share the same settings.section#models slot; enabling both throws a duplicate-registration error. This plugin is an enhanced replacement for the official models settings page — the bundled patch disables the official row (ui-settings-models / @deepseek-ai/dsh-client-ui-settings-models) automatically; if you install manually, disable that row in the plugin inventory first.

Does it conflict with other plugins?

No conflict with API proxy / relay / usage plugins.

Usage

  1. Open Settings → Models, expand your provider card (e.g. custom gpt).
  2. In Models, expand a model row (advanced area) and check the reasoning levels you want.
  3. Click Apply.
  4. Back in a conversation, select that model — the Effort menu now shows the declared levels.

Config reference (settings.yaml)

reasoningEfforts declares a model's selectable thinking levels: each key is a level selectors offer, its value the spelling dispatch sends on the wire (low: low passes the canonical name through; max: ultra renames it for a gateway with its own vocabulary). Keys come from the level set off / minimal / low / medium / high / xhigh / max; a level not declared is not offered.

llm-pi-ai:
  providers:
    gpt:
      apiKeyEnv: GPT_API_KEY
      api: openai-completions
      baseURL: https://api.example.com
      models:
        - id: gpt-5.6-sol
          name: GPT-5.6 Sol
          reasoningEfforts:
            low: low
            medium: medium
            high: high
  • off is the one three-state key: declared with no value (off:) offers Off and sends no parameter when selected; declared with a value (off: none) sends that value.
  • Once declared, the declaration is the whole offer — restate every level you want to keep.
  • For gateways with their own vocabulary, change the value (e.g. max: ultra) — no UI change needed.

License

MIT. Fork of @deepseek-ai/dsh-client-ui-settings-models; upstream copyright retained.

Comments

Loading…

From the same category

dsh-purge

by kingcharleslzy-ai

Community-maintained dsh-purge fork with host-policy patches, prompt and rule settings, diagnostics, separate update-check and install controls, and local backup/rollback.

Tools & CapabilitiesTerminal & ClientsManifest valid

★ 11.3k

MIT

JavaScript

Nov 18, 2026

dsh plugin --profile web add dsh-purge-community

by ccch1mneyyy

DSH 官方公众号收录的 TUI 补位插件:Claude Code 风,鲸鱼顶栏/实时状态/流式思考/双击 Esc 回滚/上下文进度+TPS。npm 一键装。 DSH official WeChat featured TUI plugin — Claude Code style: whale bar, live status, streaming thoughts, double-Esc rol

Terminal & ClientsManifest valid

★ 8.1k

↓ 19.2k/wk

MIT

TypeScript

Nov 22, 2026

dsh plugin --profile terminal add @deepseek-harness-tui/dsh-tui

by mmparko

DeepSeek Harness 破甲 (mirror of YuJunZhiXue/dsh-purge) — 上游即将删库,此为本仓库全量镜像:含全部分支/73 个 tag/完整提交历史。原作者署名与二开要求见 README 署名章节。

Terminal & ClientsManifest valid

★ 5.7k

↓ 14.7k/wk

MIT

JavaScript

Oct 13, 2026

dsh plugin --profile web add dsh-purge

by bowenliang123

The best DeepSeek Harness plugin for context insight and management, with context dashboard / browser / sidebar and context command, for context statistics, composition, breakdown, evolution details,

Tools & CapabilitiesUI & ExperienceTerminal & ClientsManifest valid

★ 5.4k

↓ 118.6k/wk

Apache-2.0

TypeScript

Nov 21, 2026

dsh plugin --profile web add dsh-context

by strukto-ai

The World's First Virtual Terminal for AI Agents

Terminal & ClientsManifest valid

★ 3.8k

↓ 528/wk

Apache-2.0

TypeScript

Nov 22, 2026

dsh plugin --profile agent add @struktoai/mirage-dsh

by c1overzz

DSH 插件:每天替你刷一轮 B站,读字幕挑出值得的、写一份日报,值得的 @ 你

Tools & CapabilitiesNotifications & IntegrationsTerminal & ClientsManifest valid

★ 1.5k

↓ 8.4k/wk

MIT

JavaScript

Nov 12, 2026

dsh plugin --profile web add dsh-bili-scroll