dsh-auto-vision
Run the following command in DeepSeek Harness:
dsh plugin install NormanFxxkingRockwell/dsh-auto-vision
Paste the following prompt into your AI chat to install this plugin:
To install the dsh-auto-vision plugin in DeepSeek Harness, run the command dsh plugin install NormanFxxkingRockwell/dsh-auto-vision in your terminal. Source: https://github.com/NormanFxxkingRockwell/dsh-auto-vision
About this plugin
In DeepSeek Harness, text-only main models (like DeepSeek V4 Flash) are inherently incapable of directly reading images, causing built-in read_image tools to be rejected outright, leaving the model stuck on image-related tasks. dsh-auto-vision is designed to solve exactly this pain point. It acts as a transparent visual intermediary, automatically discovering your configured multimodal models and delegating image recognition tasks to them. The plugin automatically hides the failing read_image in text-only sessions, forcing the model to take the visual route, while supporting both single and batch image recognition (up to 10) and returning results as clean text to the main model. Whether you are building agents that require vision capabilities or simply want a text-only large language model to possess "see-and-speak" abilities, this plugin is the perfect complement.
Use Cases
- Enable text-only models to recognize image content
- Automatically handle batch recognition of multiple images
- Hide failed image tools and force the visual pathway
Best For
- Developers using text-only models like DeepSeek V4 Flash
- Users needing image recognition capabilities for Agents
- Researchers wanting to add vision capabilities to models
Related Plugins
ModLens is a vision plugin for DeepSeek Harness that gives text-only models sight by reading images pasted directly into chat, with zero-config setup and multiple vision engines.
On-demand vision for text-only DeepSeek Harness agents: built-in free keyless vision chain and 14 vision tools, routing image turns as tool calls to vision models with pixel fidelity, no Python needed, one-command install.
Give text-only models in DeepSeek Harness eyes, enabling image Q&A, long-screenshot OCR, UI restoration, and GUI visual tasks.
An armor-breaking plugin for DeepSeek: rewrites instruction priority, expands refusal-blocked phrases, persists across sessions, and shows a green active indicator.