dsh-visibridge
Structured vision evidence (OCR/layout/semantics) plus a USB camera capture tool for a "shoot-look-adjust" debug loop; backends: Ollama, DeepSeek, Xiaomi.
结构化视觉证据（OCR/版面/语义）+ USB 摄像头拍摄工具，支持"拍-看-改"调试闭环；后端支持 Ollama、DeepSeek、小米。
How to install
dsh plugin add github:lhbsaa/dsh-visibridge About
dsh-visibridge **DeepSeek Harness（dsh）宿主级视觉插件** —— 为纯文本模型（DeepSeek、GLM 等）补全图像识别能力。 analyze_image 工具把图片发送给配置的视觉模型（本机 Ollama / 小米 MiMo 云端 / 任意 OpenAI 兼容端点），返回**结构化 JSON 证据**（OCR 全文、版面、场景、实体、不确定项），纯文本模型基于证据回答，不再"凭空猜测"。 宿主级 bundle 插件：**dsh 重启后自动加载**，任何会话、任何 preset 都可用 无需代理进程、无需 hook、无需修改 harness 核心配置 图片默认在本机处理（Ollama 后端零上传） --- 演示（摄像头拍摄 → 识别） 摄像头拍书本页面，capture_image 直接识别： 手机屏调试（小字 OCR）、文档扫描、视觉闭环对比见 docs/capture-image-design.md。 功能特性 | 能力 | 说明 | |------|------| | 图片识别 | 本地文件路径 / http(s) URL 均可 | | **摄像头拍摄识别** | capture_image 工具：USB 摄像头自动拍照 → 立即识别 → 结构化证据 + 图片保存；模型可在调试循环中自主调用观察变化（"改→拍→看→再改"视觉闭环） | | …
Recommendation signals
Meta
- License
- MIT
- Language
- JavaScript
- GitHub stars
- 2
- mo. downloads
- –
- Last push
- 2026-08-24
- Created
- 2026-08-17
Links
Basic safety check
- Findings
- curated 收录但无 npm 包/安装命令
- Sources
- curated:awesome-dsh-plugin.com, curated:awesome-dsh-plugin/awesome-dsh-plugin
- Topics
- camera, deepseek-harness, dsh-plugin, image-to-text, multimodal, ocr, vision, visual-loop
Related plugins
dsh-vision-router
ysr666/dsh-vision-router
Free vision for text-only agents: built-in keyless vision chain plus pixel tools (Q&A, grounding, crop, pixel diff, colors, OCR, SVG trace, cutout, screenshots); paste an image to use it.
modlens
liustack/modlens
Vision bridge for text-only models: paste an image, get structured JSON evidence (OCR, layout, semantics).
dsh-web
zhu1090093659/dsh-web-ui/tree/main/packages/dsh-tool-describe-image
A `describe_image` vision tool for text-only models: images (local path, URL, attachment) go to a configurable OpenAI-compatible vision endpoint and only the returned text enters the session.
dsh-vision-toolkit
Anionex/dsh-vision-toolkit
Vision for text-only models: paste an image and the model switches to a Vision Toolkit variant for image Q&A, multi-image comparison, long-screenshot OCR, screenshot-to-UI reproduction, element grounding, and pixel diff. No API key by default — images are processed by the author-hosted free service, 100 per machine per day; configurable to your own provider.