dsh-vision
Run the following command in DeepSeek Harness:
dsh plugin install zoahdev/dsh-vision
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install zoahdev/dsh-vision inside DeepSeek Harness to install; source code at https://github.com/zoahdev/dsh-vision.
About this plugin
DeepSeek Harness ships with a purely text-based reasoning loop. The moment a conversation needs to inspect a screenshot, a product photo, or a dashboard chart, the model has no way in. dsh-vision exists to fill that gap by giving the harness a set of eyes, so the reasoning pipeline can actually see and describe images.
Once installed, developers simply call the vision_analyze tool with a local file path or an https/data: URL and an optional question (the default prompt asks for a detailed description). The plugin forwards the request to any OpenAI-compatible vision endpoint and returns a structured text answer. apiKey, baseUrl, model, and timeoutMs are all configurable in the project settings, so you can point it at the OpenAI cloud or a local stack such as Ollama with llava, without touching any Harness core code.
It is well suited for agent workflows that require real-time image understanding: answering questions from document screenshots, describing e-commerce product photos, or interpreting dashboard metrics. It is a lightweight community plugin under the MIT license with zero runtime dependencies.
Use Cases
- Analyzing screenshots, product photos, or charts mid-conversation
- Interpreting dashboard data in real time within an agent workflow
- Running a private image-description pipeline via local Ollama and llava, no cloud dependency
Best For
- Developers who want to add vision capability to DeepSeek Harness
- Teams that prefer local deployment to keep image data on-premises
- Engineers building multimodal agent workflows
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.