dsh-read-image
Run the following command in DeepSeek Harness:
dsh plugin install OoWJZZoO/dsh-read-image
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install OoWJZZoO/dsh-read-image in your terminal and restart dsh web; the source is at https://github.com/OoWJZZoO/dsh-read-image
About this plugin
Text-only routes in DeepSeek Harness reject pasted images outright, leaving the model blind to any visual content. dsh-read-image closes that gap: it lets text routes declare image input, projects pixels into [Image #N] placeholders, and registers a first-class read_image tool backed by a configurable vision model that transcribes images into structured text—no preset edits or manual patch files required.
Key capabilities: read images by index or file path; override prompt, reasoning effort, timeout, and token budgets per call; automatically step aside on natively multimodal routes; visual settings panel with hot-reload and no restart; startup self-check that fails safe without disrupting the host.
Ideal for developers running non-multimodal text routes in agent workflows who need the model to understand pasted images, without swapping models or deploying a separate vision service.
Use Cases
- Pasting screenshots into a text-only agent conversation and getting a vision-model description back
- Reading local image files by path in automation workflows to extract text or structured output
- Swapping the vision model or tuning reasoning parameters without touching the text route config
Best For
- Developers running non-multimodal DeepSeek text routes for agent development
- Teams that need image understanding without swapping out the main text model
- Plugin users who treat a vision model as an on-demand tool in their pipeline
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.