Vision tools · DeepSeek Harness plugins — Page 7

AI image AND video generation plugin for the dsh web GUI, forked from @dickpy/dsh-imagegen. Images: text-to-image and image-to-image through configurable provider channels (gpt-image-2 / grok-imagine-image / nanobanana series / seedream-5.0-pro / dall-e-3, with native xAI Grok Imagine, Google Nano Banana and ByteDance Seedream request shaping), per-channel model catalogs, a three-column studio beside the native conversation, an infinite canvas with node-graph workflows (annotation boxes with attached prompt cards, local background removal, AI layer split, per-node model choice, file nodes with any-file upload, and a canvas skill system covering skill-driven next steps on every node: text polish, image description, content extraction, and long-running headless-agent pipelines such as image-to-editable-ppt, with a local skill library that installs and removes skills from URLs or uploaded archives), an ecommerce product-set mode and a prompt template library. Video (added by this fork): text-to-video, image-to-video, first/last-keyframe and multi-reference modes with per-channel video model catalogs, an async submit/poll/download pipeline, a clip history merged into the sidebar, favorited clips in the asset library, host-side Agent video tools, video nodes on the infinite canvas with the same controls as image nodes, and clip stitching of two or more canvas video nodes into one MP4 with the local ffmpeg in canvas reading order.

Third-party provider reasoning-effort AND input-modality settings for DeepSeek Harness: thinking levels and image-input support declared per model, auto-adapted from a model knowledge base + wire-protocol inference, edited right inside the official Models page card.

基于deepseek harness的短剧/动画画布工作室：画布可挂会话视图页或右侧栏（边看对话边生成，由 canvasHome 配置）+ ComfyUI 生成服务（图片 FLUX / 视频 MiniMax H3 音视频，带声音/参考绑定/原生字幕/fast·balanced·quality 三档）+ Agent 画布工具 + 自带 skill。双面 DSH 插件。

Sync active Wallpaper Engine wallpapers(image/video/scence/app)to DeepSeek Harness Web UI as a customizable dynamic background.

DirectorX：第一个以 DeepSeek Harness 为内核的 Video Agent Harness——面向端到端 AI 视频生产的开源 dsh-plugin，覆盖视频生成、分镜画布、智能剪辑、FFmpeg 质检与交付。The first DeepSeek Harness-native video agent harness for AI video production.

DeepSeek Harness dual-face plugin: Codex-style file references (@path), fully bundled document→Markdown (MarkItDown engine, 20+ formats), mature image handling (read_image for multimodal routes, automatic vision descriptions for text-only models), read_document tool.

A desktop companion for DeepSeek Harness: its own persona, lorebook and multimodal API, a user profile distilled from the harness's own chat history, awareness of the running task, proactive chatter, reminders, voice, 8-24 frame sprite animations, screen reading and computer control behind a three-level permission, and a click-to-talk desktop window (WinForms sprite or Edge app window) next to the in-app floating pet.

Drop any file onto the DeepSeek Harness chat: non-image files are saved into the session workspace's .dsh-uploads/ folder and referenced from the composer as @.dsh-uploads/<name>, so the agent reads them with its file tools. Images keep going to the core's own attachment rail.

Extra media APIs for DeepSeek Harness: image generation, video generation, text-to-speech and transcription behind one settings panel, with adapters for OpenAI-compatible endpoints, Google Gemini/Veo, Replicate, fal, ElevenLabs, Fish Audio, MiniMax and any custom HTTP endpoint.

Skin Studio for DeepSeek Harness: a model-facing skin_studio tool that lets the agent design and one-click install launcher or frontend skins on user request — with a guided flow (ask requirements first; request an image from the user when the model cannot generate one, or call an image-generation tool when it can), local image inlining as data URIs, and instant apply through the DSH Launcher API.

图片理解桥接(dsh-vision-bridge 的中文界面分支):聊天里的图片交给你选定的视觉模型描述,非视觉模型也能"看图"。Fork of @goodandready/dsh-vision-bridge 0.4.5 (MIT) with a Chinese UI; the wire, routes and config file are unchanged.

Aggregate cordis bundle for the dsh-ros2 plugin family — depends on core/profile/moveit/safety/vision/common; empty apply for backward compatibility (install this to get all 83 tools + 9 skills).

dsh-ros2 realtime vision: image-topic acquisition, parallel VLM analysis, vision pipeline tools + vision provider service (7 tools) + the vlm/ and offscreen/ ROS2 packages (published) + robot-state-vision-analysis skill. Cordis bundle.

DeepSeek Harness QoL plugin for the local Qwen3.8 line (27B, Flash-Next): per-request thinking budgets on llama.cpp, NInfer, TabbyAPI, and oMLX dialects, vision + tools, and a compaction backend whose summaries stop burning the output cap on thinking.

A second brain for your DeepSeek Harness agent — strong-model review before risky tool calls, plus 75 models from one credential: a BlockRun API key, or a wallet paying per call over x402.

Windows computer use for DeepSeek Harness: an MCP stdio server + PowerShell UIA backend, bridged into DSH via @deepseek-ai/dsh-mcp-client. Read/act on real Windows desktop apps (UI Automation tree, screenshots, typed input, OCR, window management).

Find the doc your agent just wrote. Lists the DSH workspace's Markdown, images and video in the sidebar, ranked by relevance to the current conversation, with inline preview — and hands the same ranking to the agent as a tool, returning the passage that matched in each document. Zero model calls, zero network.

Appearance customization plugin for DeepSeek Harness WebUI: theme palette, background image, transparency and blur over the --dsw-* token system

Free image reading and generation skills for DeepSeek Harness v0.1.6: visual QA, illustration generation, provider fallback, and no bundled keys.

DSH theme plugin for the DeepSeek Harness Web UI: Nous Blue palette, ambient glow, dynamic image/video backgrounds. DSH 界面设定插件：配色 / 氛围光 / 动态背景（中英双语）

webui：视图图块/消息导航 + 工具调用聚合 + Markdown 渲染 + 模型推理等级同步 + AnySearch 网页搜索（合并 dsh-tool-summary / dsh-better-markdown / dsh-web-search-anysearch）

WebUI 文件列表与选中文件预览：输入框工具行按钮打开右侧浮层，浏览目录、预览文本与图片，右键文件可插入官方 @ 引用/路径/内容到输入框，不修改任何官方文件

WebUI 右侧侧边栏：① 文件资源管理器（工作区根 sandboxPolicy.workspaceRoot 为默认根、目录优先/隐藏文件灰显、支持上级任意路径浏览）；② 文件预览/编辑（受控文本区 + 图片/PDF/MD 预览、保存回写）；③ 内嵌浏览器（沙箱 iframe + 地址栏导航）；④ CMD 终端（轻量 SSE 流 + cmd.exe 逐行执行）；⑤ 任务管理（后台任务列表/输出/停止）。零原生依赖、零构建，纯 JS 注入。参考/复刻自第三方插件 DSH Better Sidebar（omdsh-dev/DSH-better-sidebar），本版内置其资源管理/文件预览/浏览器/终端/任务五项核心能力。

数字地球：在三维地球上叠加 3D Tiles / 3DGS 图层，与对话同屏；Agent 可读写场景、标注、量测、空间分析与视觉特效，支持 /earth 命令、问 Agent 场景注入与 HTML/Markdown 报告导出，内置数据体检、多期场景恢复与双语界面。