dsh-image-plugins
Run the following command in DeepSeek Harness:
dsh plugin install alanzhao0128/dsh-image-plugins
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install alanzhao0128/dsh-image-plugins in your DeepSeek Harness terminal to install this plugin; the full source is at https://github.com/alanzhao0128/dsh-image-plugins . Restart dsh web after installation to activate it.
About this plugin
the DeepSeek official text-only route cannot accept or emit images, yet everyday workflows often require the main model to see or create pictures. dsh-image-plugins bridges this gap with two model tools - understand_image (vision) and generate_image (generation) - letting a text-only main model delegate multimodal tasks to an external endpoint without switching its routing model.
Both tools accept any user-supplied endpoint: any OpenAI-compatible API (OpenAI, SiliconFlow, Zhipu, Tongyi compatible mode, Ollama, etc.) or Alibaba DashScope native Model Studio interface. The vision tool writes the model text description into the session log so the main model can reason over it; the generation tool saves the result into the workspace and returns the path; with the DashScope provider an optional reference image enables image-to-image editing. Every capability is opt-in; an unconfigured install is completely inert and ships no keys.
Built for developers whose main model runs on a DeepSeek text route and who also need third-party vision or image-generation APIs, or for teams that want to drive look-at-image, generate-image, and edit-image flows naturally inside the dsh workspace while keeping endpoint credentials centralized and out of the session log.
Use Cases
- Read a workspace screenshot and get a text description via natural language
- Generate an image from a prompt and save it into the workspace automatically
- Edit an existing image in place using a reference image via DashScope
Best For
- Developers running a DeepSeek text-only route as the main model
- Teams integrating third-party vision or image-generation APIs
- Developers who want centralized multimodal credential management without leaking keys into logs
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.