플러그인
DeepSeek-Harness 플러그인을 찾아보고, 필터링하고, 설치하세요.
플러그인 397개 찾음
dsh-lcx-codex
kk3ya03-star/dsh-lcx-codex
Sub2API 프록시 또는 NewAPI 중계 GPT 모델용 DSH 웹 검색 및 Responses 네이티브 V2 컴팩션
dsh-audio-copilot
ai-yucheng/dsh-audio-copilot
DeepSeek Harness용 오디오 코파일럿: 오디오 전사(ASR) 및 음성 합성(TTS) — 텍스트 전용 에이전트에 귀와 목소리를 제공합니다. Windows 로컬 SAPI TTS 즉시 사용 가능; OpenAI 호환 ASR/TTS 엔드포인트 구성 가능. 컴포저 내 음성 입력 포함
dsh-plugins
huyang218/dsh-plugins
서드파티 OpenAI 스타일 게이트웨이를 위한 DeepSeek Harness 호환성 플러그인: 완전한 응답 후 [DONE] 센티넬 없이 닫히는 SSE 스트림을 허용합니다.
dsh-hub-oauth-gateway
lninghaha/dsh-hub-oauth-gateway
로컬 SQLite 기록, HUD 및 대시보드, Hub 계정/할당량 스냅샷, 사용자 지정 가격 책정 및 요금 장부, CSV/JSON 내보내기를 갖춘 DSH 웹 사용 센터와 코딩 플랜 OAuth 및 선택적 루프백 OpenAI/Anthropic 게이트웨이.
dsh-image-plugins
alanzhao0128/dsh-image-plugins
DeepSeek Harness용 멀티모달 플러그인으로, 구성 가능한 OpenAI 호환 또는 DashScope 엔드포인트를 통해 이미지를 이해하고 이미지를 생성합니다.
dsh-plugin-dit
cuboteam/dsh-plugin-dit
DeepSeek Harness용 DIT.ai 제공자 번들: 하나의 DIT_API_KEY에서 29개의 OpenAI Chat Completions 모델과 10개의 Anthropic Messages 모델을 설치하며, 프로토콜별 추론 구성을 제공합니다.
dsh-tool-vision
wanshichenguang/dsh-tool-vision
DashScope OpenAI 호환 API(qwen3.7-flash) 기반의 모델용 image_describe(识图) 도구와 붙여넣기 브리지: 텍스트 전용 세션에서 붙여넣은 이미지가 전송 시 파일 경로로 자동 변환되고 트랜스크립트에 다시 렌더링되어 이미지 허용 단계에서 걸리지 않습니다. DASHSCOPE_API_KEY는 BYOK, 엔드포인트/모델/예산 설정 가능, 리다이렉트 방지 HTTP 클라이언트, 모든 agent preset에서 동작.
dsh-eyes
leeminjing/dsh-eyes
텍스트 전용 DeepSeek 모델용 온디맨드 비전: 이미지를 업로드하면 모델이 임의의 OpenAI 호환 비전 엔드포인트(기본 Qwen/DashScope)를 사용하는 view_image 도구를 호출합니다.
dsh-tool-vision
gloryxpnv/dsh-tool-vision
텍스트 전용 에이전트를 위한 로컬 우선 구조적 비전: 이미지를 로컬 OpenAI 호환 VLM에 보내고 JSON 증거(요약, 그대로의 OCR, 레이아웃 영역, 엔티티/관계, 색상, 명시적 불확실성)로 돌려받습니다. 환각 방지 폴백과 선택적 붙여넣기/업로드 브리지 포함. 클라우드 비용 없음, 이미지는 기기 밖으로 나가지 않습니다.
dsh-vision-fallback
1helloman1/dsh-vision-fallback
채팅 이미지를 고정된 OpenAI 호환 비전 모델로 라우팅하고 사실에 근거한 관찰을 선택된 메인 모델에 반환합니다. 세션 범위 관찰은 리플레이·압축·재시작을 넘어 재사용됩니다.
model-catalog
johnxu22786/model-catalog
OpenAI 호환 API 호스트에서 모델 목록, 가격, 기능을 자동으로 발견해 정규화하고, dsh용으로 바로 쓸 수 있는 모델 구성을 출력합니다.
dsh-api-balance
gpiox/dsh-api-balance
DeepSeek Harness용 플로팅 API 잔액 배지: DeepSeek, Moonshot, OpenAI 및 사용자 정의 엔드포인트의 실시간 잔액; 드래그 및 크기 조절 가능, 프로스트 글래스 스타일, 아래 콘텐츠에 맞춰 변하는 텍스트 색상.
dsh-plugin-thinking-mode
lovedheart/dsh-plugin-thinking-mode
DeepSeek Harness 플러그인: OpenAI 호환 엔드포인트의 Qwen3 계열 모델을 위한 요청별 사고 모드 전환(enable_thinking)
dsh-jev-harness
zrxparley/dsh-jev-harness
One switch decides who does the work in DeepSeek Harness. OFF steps aside and uses the DSH model gateway; ON sends every instruction to a JEV judge first, which routes judgment tasks to Kimi and routine execution to DeepSeek Flash. By default the routed calls run through the DSH model gateway (ctx.llm), so only the JEV judge key is needed here; the strategy and execution models are picked from the providers and model ids this DSH install actually has (ctx.llm.listProviders/listModels), auto-ranked or pinned in the panel; Direct mode calls your own OpenAI-compatible gateway instead. Includes a left-sidebar settings page, a composer switch, a hover decision trace (which model actually answered), a traffic dashboard, and a skill router that asks the same judge which of the 39 mattpocock/skills fits the task. Tools: harness_call/harness_classify/harness_review/harness_probe/harness_skill/harness_dsh. No user-facing tier choice. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-local-model-worker
cloudylapis/dsh-local-model-worker
Starts a local OpenAI-compatible model server only when the model is about to be used and unloads it after an idle period, with presets for LM Studio, Ollama and llama.cpp, custom start/stop commands for anything else, and a local_model tool for status, start and stop. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-drawcraft
vhow123/dsh-drawcraft
Drawcraft — an image-generation studio in the sidebar panel list, split into three pages that share one result. Prompt page: pick a target dialect, a style and an aspect, write one rough sentence, and a chat model returns an 11-field shot breakdown that the plugin stitches into the syntax that dialect actually accepts — six dialects (natural English, natural Chinese, GPT-Image, Midjourney, Stable Diffusion, nano-banana), 18 style presets, negative prompts, filler-word stripping and local history; an image can also be reverse-engineered into the same fields, and a one-line edit such as light turns to dusk rewrites only the named fields. Image page: text-to-image and image-to-image through a locally installed image-generation plugin or through an OpenAI-compatible endpoint you add in the panel, whose key stays on disk and is never echoed back to the browser; with a subject image attached it posts to /images/edits and carries up to four extra references. Canvas page: an infinite canvas where image, prompt and result nodes are dragged, resized, wired by hand or by generation, with undo and redo, marquee selection, copy and paste, a minimap, pointer-centred wheel zoom and JSON export and import; rendering from a node walks the whole upstream chain so no reference is dropped, one node can be pinned as the subject, and the reference library opens inside the canvas. A reference library of 4,345 Chinese-labelled entries — 1,439 poses with matching depth maps, 1,407 outfits, 1,499 portraits — is searched and faceted by pose system, angle, category, season, hair length and hairstyle, paged 24 at a time, and the entries you pick are injected into the prompt as binding constraints. Renders can pick the image channel and model, one of eight aspect sizes or follow-the-aspect, and quality at 1K/2K/4K; a preview opens into a lightbox with pointer-centred wheel zoom from 20% to 800%, drag-to-pan, double-click reset and one-click save. The prompt tools and the library are also exposed as the optimize_image_prompt, patch_image_prompt and search_reference_gallery agent tools. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-image-gen
adithyanraj03/dsh-image-gen
Multi-provider image generation and editing for DeepSeek Harness (OpenAI-compatible relays with custom /v1/images/generations endpoints, Gemini, Seedream, DashScope, Grok, GLM, local ComfyUI, subscription accounts), an Excalidraw MIT infinite-canvas studio with click-to-place and selection-to-edit references, and a pro viewer with zoom/pan/pinch, region notes and a markup mode that bakes marks into the pixels and sends normalized coordinates to the model.
dsh-connect-officeAce
hxgh776/dsh-connect-officeace
Connects a locally signed-in Huawei OfficeAce desktop account to DSH as an OpenAI-compatible chat provider that reuses the stored login instead of an API key, with streaming, reasoning and tool-call passthrough.
dsh-asr-openai
caiccc-a11y/dsh-asr-openai
可以直接本地调用DSH内置语转文模型的插件,支持openai标准格式。
dsh-plugin-litsearch
peijiping/dsh-plugin-litsearch
Model-driven, multi-round literature search in chat: the sidebar visualizes results, rounds and full-text availability, across ten free sources — PubMed, Europe PMC, OpenAlex, Crossref, Semantic Scholar, arXiv, PubTator, DOAJ, OpenReview and OpenAIRE — with PDF download, citation export and Zotero from either side.
dsh-xdpool
etosurvivor/dsh-xdpool
Heybash model pool for DeepSeek Harness: account rotation, provider aggregation, routing chains, usage insights and a local OpenAI-compatible gateway.
feiyu-memgate (dsh-feiyu-memgate)
n0rmal-114/feiyu-memgate
Long-term memory through the bundled feiyu-memgate gateway, which the plugin starts on 127.0.0.1 (needs Python 3.11 or newer, nothing to pip install): recalls related memories at the start of each user turn and appends them after the user message as a marked data block, gives the model memgate_remember and memgate_search tools, and adds a /memgate command and a settings page where each feature can be switched off. By default memories are written only when the model calls the remember tool. Keyword retrieval over SQLite by default, with no model and no download; an OpenAI-compatible embeddings endpoint or the optional dsh-feiyu-memgate-embed add-on (local CPU embedding models, downloaded on request) adds semantic retrieval. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
open-harness-plugin-openai-compatible-auto
eliotstocker/open-harness-plugin-openai-compatible-auto
Multi-host plugin (OpenCode V2, Pi, DeepSeek Harness) for OpenAI-compatible endpoints with /models auto-detection and LiteLLM browser CLI code flow sign-in
dsh-imagegen-panel
gpta76/dsh-imagegen-panel
Image-generation sidebar panel for DeepSeek Harness: works with any OpenAI-compatible image endpoint, supports reference-image edits, and ships presets for character sheets, turnarounds and expressions.