DeepSeek Harness Plugin Hub

Publish and manage complete Harness Profiles. Discover Plugins for your next setup.

Explore

PluginsPresetsDocsNews

Community

Publish a pluginContactReport an issue

Resources

Plugin Hub on GitHubDeepSeek HarnessSystem statusPrivacy notice
© 2026 DeepSeek Harness Plugin HubPowered byPaxTech

Independent and unofficial. Not affiliated with, authorized by, or endorsed by DeepSeek.

CATEGORY

Vision & Media DSH plugins

Verified manifests and exact versions in this category.

All categories

All pluginsAgents & Orchestration1.5kMemory & Context2.1kCoding & Developer Tools4kUI & Customization5.6kIntegrations & Communication3kVision & Media919Search & Research1.2kSecurity & Access1.8kModels & Usage2.5kProductivity & Workflow3.6k

All plugins

912 plugins
Pv0.6.1
dsh-plugin-voz-copilot

Plugin Voz Copilot

Voice copilot for DeepSeek Harness: names and attaches the push-to-talk session, adds composer microphones, and runs a real-time spoken conversation loop with Silero VAD, speculative transcription, streaming TTS and speaker verification.

↓ 0★ 0
Updated 10/18/2026
Dv0.6.4
dsh-douyin-publish

Douyin Publish

Take a local video all the way to a creator post on Douyin or Bilibili: optional background music mixed in with ffmpeg before upload, self-check the page layout, hand the file straight to the page's own upload control over a loopback one-file server (no native file dialog, falls back to it for huge

↓ 0★ 0
Updated 10/18/2026
Ov1.0.0
@local/dsh-opening

Opening

DSH opening film: plays the 19.5s monochrome title sequence on startup, with skip, sound and replay.

↓ 0★ 0
Updated 10/17/2026
Vv0.1.0
dsh-voice-live

Voice Live

Real-time duplex voice for DeepSeek Harness: Volcengine streaming ASR/TTS, agent reply narration, barge-in, wake word, live captions. | 实时双工语音插件：火山流式 ASR/TTS、回复朗读、打断、唤醒词、实时字幕。

↓ 0★ 0
Updated 10/17/2026
Iv0.4.1
dsh-image2-tool

Image2 Tool

DeepSeek Harness bundle that exposes an OpenAI-compatible image relay as the image2_generate tool for text-to-image and image-to-image, with inline media previews.

↓ 0★ 0
Updated 10/17/2026
Rv0.1.0
read-figure

Read Figure

Give text-only models a fixed three-section figure-reading handoff: a vision model describes a physics figure (structure / restatement / confidence), the tool returns it as text, and uncertain marks must be listed instead of guessed.

↓ 0★ 0
Updated 10/17/2026
Bv1.0.0
dsh-boot-animation

Boot Animation

Play a full-screen video over the DeepSeek Harness boot screen and hand over to the shell when it ends.

↓ 0★ 0
Updated 10/16/2026
Qv1.0.0
dsh-quick-screenshot

Quick Screenshot

One-click DSH screenshot: capture the DeepSeek Harness window (or the whole screen) from the Host and drop the PNG straight into the composer draft

↓ 0★ 0
Updated 10/16/2026
v0.1.0
dsh-custom-vision

Custom Vision

Use your own OpenAI-compatible vision model in DeepSeek Harness: images are described before the main model sees the request, plus an image_understand tool for reading charts, screenshots, UI and OCR.

↓ 0★ 0
Updated 10/16/2026
Mv0.1.0
dsh-media-sidebar

Media Sidebar

Right-sidebar media tab for DeepSeek Harness: inline playback for local audio/video — HTTP Range for native formats, ffmpeg transcoding for mkv/avi/flv/wmv.

↓ 0★ 0
Updated 10/16/2026
Vv1.0.0
dsh-video-clone

Video Clone

把一支视频复制成 N 份，每份塞一个唯一 GUID 让文件哈希互异，而媒体流与解码帧逐字节不变。Copy one video into N files, each carrying a unique GUID so every file hash differs, while the media stream and decoded frames stay byte-identical.

↓ 0★ 0
Updated 10/16/2026
Vv0.1.1
dsh-video-grab

Video Grab

Video inspection for DSH agents: probe container structure, extract frames, build one contact sheet that shows a whole clip in a single image, and transcribe speech locally. Works without ffmpeg by driving the bundled Electron/Chromium decoder.

↓ 0★ 0
Updated 10/15/2026
Cv0.1.0
dsh-crop-zoom

Crop Zoom

Magnify a region of a screenshot so small text becomes readable: a DeepSeek Harness (dsh) tool plus a standalone CLI that crops and upscales before a model reads it.

↓ 0★ 0
Updated 10/15/2026
v0.1.14
dsh-voice-announcer

Voice Announcer

End-of-conversation voice announcement: session name + turn count + result (edge-tts streaming / SAPI)

↓ 0★ 0
Updated 10/15/2026
Rv0.1.2
dsh-read-animation

Read Animation

Let a DeepSeek Harness agent actually see an animated GIF or animated WebP: decode every frame, sample the animation across time, and hand the model a labelled sequence of real frames instead of the single first frame the harness otherwise delivers.

↓ 0★ 0
Updated 10/15/2026
Mv0.1.1
dsh-minimax-tts

Minimax Tts

DSH plugin: Use MiniMax（海螺）TTS to read assistant messages in the conversation stream aloud—manual playback + Dynamic Mode (Agent registers scripts for reading aloud)

↓ 0★ 0
Updated 10/14/2026
Vv1.0.0
dsh-voice-doubao

Voice Doubao

Adds two Volcengine Doubao cloud speech recognizers to the DeepSeek Harness voice-input picker, optionally rewriting the transcript with an LLM. / 给 DeepSeek Harness 的语音输入加入两个豆包云端识别器，可选再用大模型把口语顺成通顺的书面句子。

↓ 0★ 0
Updated 10/14/2026
Vv1.0.0
dsh-vision-unlock

Vision Unlock

Remove the restriction that causes DSH to refuse sending or reading images because the model has not declared multimodal capabilities.

↓ 0★ 0
Updated 10/14/2026
Vv0.1.1
dsh-vision-reasoner

Vision Reasoner

Mark things on a picture and look at the picture: a DeepSeek Harness plugin that turns an image into a record of numbered marks, with coordinate-labelled close-ups and comparison sheets you can check.

↓ 0★ 0
Updated 10/14/2026
v0.2.1
@vibedev-si/dsh-media-viewer

Media Viewer

右侧栏媒体预览：直接预览视频、音频和 HTML（大文件可拖进度条，HTML 默认允许脚本），文件夹媒体画廊（多选后加入影视工作台画布）；用 DSH 自带的右侧栏，装了 dsh-better-sidebar 时也在其中可用 · Inline video, audio and HTML preview plus a folder media gallery in the host's own right sidebar (and inside dsh-better-sidebar when installed) — range-streamed large files, sandboxed HTM

↓ 0★ 0
Updated 10/14/2026
Iv0.4.0
dsh-image-compare

Image Compare

Sidebar image comparer for DeepSeek Harness: browse the session workspace inside the right-column panel, open a set of images into one pageable viewer, zoom and pan each image independently, sync one view onto the whole set, and reorder them by dragging the filmstrip.

↓ 0★ 0
Updated 10/14/2026
Rv0.1.0
dsh-readvideo

Readvideo

DeepSeek Harness plugin: Read videos by fetching bytes on demand (Range seeking + frame extraction + speech transcription), without downloading the entire video

↓ 0
Updated 10/14/2026
Rv0.1.0
dsh-read-video

Read Video

DeepSeek Harness plugin: read videos by fetching bytes on demand (Range seeking + frame extraction + speech transcription), without downloading the entire video

↓ 0
Updated 10/14/2026
Ov0.1.0
dsh-omni-modal

Omni Modal

Adds an "Omni-modal settings" section to DeepSeek Harness: configure OpenAI-compatible image / video generation endpoints with switchable config profiles, and exposes generate_image / generate_video / check_video tools to the model.

↓ 0★ 0
Updated 10/13/2026
Wv1.0.0
dsh-webp-relabel

Webp Relabel

Relabels DSH's normalized WebP request images as PNG so local OpenAI-compatible vision servers accept them.

↓ 0★ 0
Updated 10/12/2026
Sv1.0.6
@deepwhale-cn/dsh-shell-canvas

Shell Canvas

Turn a piece of text into a video ready to publish. Paste a script, choose a few assets, and click once to create both landscape and portrait versions; the script, assets, and finished videos all stay on your own computer—no uploading or internet connection. The free version is fully usable, with th

↓ 0★ 0
Updated 10/12/2026
Vv0.2.35
vision

Vision

Vision DSH bundle: give the agent screen/window vision — see tool captures via the bundled Python cvision and returns the image natively

↓ 0★ 0
Updated 10/11/2026
Cv0.2.0
dsh-creative-plugin

Creative Plugin

DSH creative mode plugin: registers a creative_generate_image tool that bridges DeepSeek Harness sessions to any OpenAI-compatible text-to-image endpoint (default: StepFun step-image-edit-2).

↓ 0★ 0
Updated 10/11/2026
Vv0.2.0
dsh-video-transcribe

Video Transcribe

DeepSeek Harness plugin: Paste a video link to download the audio track with a local toolchain and transcribe it offline (yt-dlp + ffmpeg + sherpa-onnx/SenseVoice), with automatic chunking for long audio and no cloud APIs.

↓ 0★ 0
Updated 10/11/2026
Yv0.2.0
dsh-youtube-study-notes

Youtube Study Notes

Turn a YouTube link into study notes: pulls the video's own captions, optionally reads on-screen content with a vision model, then asks for your own notes and merges everything into one document.

↓ 0★ 0
Updated 10/10/2026
← Previous1…2122232425…31Next →
DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in