AI Agent Hub
Back to plugins
🤖

dsh-image-mmx

Model Inference Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install fengs2021/dsh-image-mmx

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

To install this plugin in DeepSeek Harness, please execute dsh plugin install fengs2021/dsh-image-mmx in your terminal. Source: https://github.com/fengs2021/dsh-image-mmx

About this plugin

Traditional text-only Large Language Models (such as DeepSeek) often struggle to process visual inputs, leading to errors when users attempt to upload images. The dsh-image-mmx plugin is designed to solve this pain point by giving text models a pair of "eyes." When a user sends an image in a chat, the plugin automatically saves it locally and invokes the local MMX (MiniMax VLM) for visual recognition, extracting key information such as text, layout, and details from the screen. This recognized text is then seamlessly injected into the model's context, allowing text-only models to understand image content. Supporting parallel processing for multiple images and gracefully degrading to file paths if MMX is unavailable, this tool is ideal for developers who need to handle charts, code screenshots, or UI elements.

Use Cases

  • Processing code screenshots or technical docs
  • Analyzing software interfaces or charts
  • Uploading images for intelligent parsing

Best For

  • Developers using text-only models like DeepSeek
  • Users needing to extract data from visuals
  • AI workers wishing to add visual capabilities to text models