AI Agent Hub
Back to plugins
🤖

dsh-auto-vision

Model Inference Updated 2026.08.26

Run the following command in DeepSeek Harness:

dsh plugin install soarGuo/dsh-auto-vision

Paste the following prompt into your AI chat to install this plugin:

In DeepSeek Harness, run the command dsh plugin install soarGuo/dsh-auto-vision to install the dsh-auto-vision plugin; the full source code is available at https://github.com/soarGuo/dsh-auto-vision.

About this plugin

In AI conversations, sharing images is common, but many models like deepseek-v4-pro lack image input support, preventing users from sending screenshots directly.

The dsh-auto-vision plugin addresses this issue. It automatically recognizes images using a configured vision model and writes descriptions as folded context entries into the session history. The original message remains intact, allowing the model to indirectly understand the image content.

This plugin is ideal for all DSH users who need to handle images, especially developers or researchers relying on models without native image support. It's implemented purely as a plugin without modifying DSH source code, making it easy to install and use.

Use Cases

  • Automatically recognize images when sending screenshots to models without image input support.
  • Add image descriptions as folded entries in the session while preserving the original message.
  • Bridge image processing with vision models to enhance interaction efficiency.

Best For

  • Developers using DSH for model interactions.
  • Researchers who need to process image data.
  • General users wanting to share screenshots when models lack support.