dsh-imgdraw
Run the following command in DeepSeek Harness:
dsh plugin install NinjaSln-labs/dsh-imgdraw
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install NinjaSln-labs/dsh-imgdraw in your DeepSeek Harness terminal to install the plugin; the source lives at https://github.com/NinjaSln-labs/dsh-imgdraw and it auto-loads with your profile on every restart.
About this plugin
Generating avatars, illustrations, or visual assets is a frequent need in everyday DeepSeek Harness chats, yet the usual workflow forces you to leave the conversation, operate in a separate tool, and paste results back—a disjointed loop that breaks context. dsh-imgdraw eliminates that friction by embedding text-to-image directly into the Harness workspace: a single button beside the input opens a modal where you type a prompt, pick an aspect ratio, choose how many images to generate, switch backends, and watch a live quota counter. Results appear in a four-grid preview where each tile can be downloaded, flagged as permanent, or deleted, and the full history survives restarts.
Three usage paths power the plugin. The interactive modal handles batch generation of one to four images across five aspect-ratio presets, includes one-tap avatar templates, and displays real-time quota feedback. You can switch between DashScope and SiliconFlow backends without touching any code. The draw_image model tool lets the AI call image generation autonomously inside a conversation, waiting synchronously (up to 180 seconds) and returning direct image URLs—ideal for letting the model produce visual content as part of a larger task. The /imgdraw/ static route serves generated images as local binary files on loopback, so they can be hot-linked into pages or documents.
The default backend is DashScope wan2.7-image on its free tier; switching to SiliconFlow Qwen-Image, or falling back to qwen-image-2.0 and z-image-turbo once the free quota is exhausted, is a single config change. The plugin ships as a formal bundle that auto-loads with your profile and survives restarts—no re-approval, no lost state.
It suits developers and content creators who generate images regularly within Harness, want the entire generation loop to stay in the same pane as the conversation, or prefer letting the model invoke image tools on its own. All files stay on your machine, quotas are tracked locally with a configurable history cap, and there is no external account dependency—just drop in an API key and start drawing.
Use Cases
- Generate avatars or illustrations directly inside Harness chats without leaving the window
- Let the AI invoke the draw_image tool autonomously to produce visual assets in multi-step tasks
- Batch-generate multi-ratio images and manage local history with a keep-or-delete policy
Best For
- Developers who frequently need image output within Harness
- Content creators who want image generation in the same pane as the chat
- Automation engineers who prefer the AI invoking image tools on its own
Related Plugins
A unified suite combining hot runtime injection, task-aware thinking-mode routing, and a graded session protocol with red-team gates to sustain model diligence across long-horizon inference.
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.