dsh-multimodal

Verified

给 DeepSeek 安装一双眼睛和一支画笔:会话里直接贴截图/图片,GLM 视觉模型先精确转写图片内容(报错信息、代码、界面逐字保留),然后 DeepSeek 继续处理你的问题——同一轮完成,全程无感;需要配图时,DeepSeek 自动调用文生图后端出图并显示在会话中。

View on GitHub

Verified plugin

Depends on the Cordis runtime or ships a cordis.yml, so it loads as a real dsh plugin.

What we found in the repo

— depends on @deepseek-ai/cordis
— depends on @deepseek-ai/dsh-llm
— npm package dsh-multimodal

Installing

Install the package, then register it in your cordis.yml:

npm install dsh-multimodal

# cordis.yml
plugins:
  dsh-multimodal:

Plugin keys and config options vary — check the repo's README before copying this in. dsh is in developer preview and its loader format is still changing.

Repository details

Stars: 1
Forks: 0
Open issues: 0
Language: TypeScript
License: MIT
npm package: dsh-multimodal @ 0.2.0
Created: 2026-08-13
Last push: 2026-08-13

Topics

deepseek-harnessdshdsh-plugindsh-pluginsmultimodalmultimodal-aivision

More in Vision & Media

dsh-vision-router

Verified

ysr666

Eyes for text-only DeepSeek Harness agents: built-in free vision chain (no key) + pixel-level vision tools (Q&A, grounding, crop, pixel diff, colors, OCR, SVG trace, cutout, screenshots). One-command install, no Python, image turns work like ordinary tool-calling turns.

17JavaScriptupdated today

dsh-vision

Verified

william-jin-cmu

dsh 插件：给纯文本 DeepSeek 加视觉——view_image 工具桥接任意 OpenAI 兼容 VLM（默认智谱免费档，实测 4 厂商 10 模型）

15TypeScriptupdated yesterday

dsh-plugin-mineru

Verified

HuanLinOTO

向模型暴露 MinerU 文档解析工具，将 PDF/图片/DOCX/PPTX/XLSX 转为结构化 Markdown/JSON | Exposes MinerU document-parsing tools to the model, converting PDF/images/DOCX/PPTX/XLSX into structured Markdown/JSON

7TypeScriptupdated today

dsh-vision

Verified

oil-oil

Near-native image understanding for DeepSeek Harness

7TypeScriptupdated today

dsh-openmaic

Verified

THU-MAIC

OpenMAIC for DeepSeek Harness: classrooms, slides, interactive widgets, and Socratic teaching

5JavaScriptupdated today

dsh-plugin-deepeye

Verified

Favio8

DeepEye vision plugin for DeepSeek Harness (DSH): image description, OCR, VQA, UI layout, and clipboard analysis.

4TypeScriptupdated today