dsh-mindsee
基于 MindSee 的 DeepSeek 图像能力插件,使 DeepSeek 能处理截图与图片输入。
AI 分析
适合图片理解、截图解读需求。需在 MindSee 后端配置 token,并重启 dsh web。
安裝
$
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:123cdxcc/dsh-mindsee說明文件
閱讀完整 README ↗MindSee
DeepSeek cannot see pictures. MindSee gives it a way to look.
Paste a screenshot into the composer, or point at a JPEG, PNG, or WebP on your machine. The image is described through MindSee; DeepSeek answers from that description. You stay on the same model—there is no need to switch to a vision model.
Install
You will need DeepSeek Harness and the web profile.
dsh plugin --profile web add dsh-mindsee
Restart dsh web when it finishes.
Connect your token
- Create an access token at mindsee.app.
- In DeepSeek Harness, open Settings → Plugins → MindSee.
- Paste the token and save.
Ask with a picture
Paste an image into the composer and write your question in the same message. JPEG, PNG, and WebP are supported, up to 20 MB.
A local file path works just as well, if you would rather type than paste.