AI Agent Hub
Back to plugins
dsh-chat-imagine preview

dsh-chat-imagine

Web Tools Updated 2026.08.26

Run the following command in DeepSeek Harness:

dsh plugin install corrinehu/dsh-chat-imagine

Paste the following prompt into your AI chat to install this plugin:

You can install this plugin in DeepSeek Harness by running `dsh plugin install corrinehu/dsh-chat-imagine`; the full source code is available at https://github.com/corrinehu/dsh-chat-imagine.

About this plugin

Generating or understanding images directly within the DeepSeek Harness (DSH) chat often requires cumbersome tool switching or complex model routing. The dsh-chat-imagine plugin seamlessly solves this by integrating image generation and visual recognition directly into your chat window. Whether you prefer using OpenAI-compatible APIs or leveraging local CLIs (like mmx, codex, or agy) to utilize your existing account quotas, the plugin automatically detects available channels and displays the generated images inline.

Beyond its robust image generation capabilities, the plugin's image recognition feature is equally impressive. It harnesses the vision capabilities of local CLIs to convert images into structured JSON evidence, including OCR text, layout regions, and semantic relationships. This allows even pure text models to perform deep analysis without needing to switch to a vision model. For DSH users seeking an efficient workflow, wanting to maximize local CLI quotas, or needing to empower text models with visual understanding, this plugin is an indispensable productivity tool.

Screenshots

Use Cases

  • Generate and inline-preview images directly in the chat window using text prompts.
  • Provide structured image recognition evidence for text models using local CLI quotas.
  • Automatically fallback and switch between different generation or vision backends to handle account quota limits.

Best For

  • DSH users who frequently need to generate and view images within conversations.
  • Developers looking to leverage existing CLI account quotas instead of relying solely on API keys.
  • Researchers or engineers who need to empower pure text LLMs with image understanding and analysis capabilities.