What problem it solves

Image creation often requires switching between separate tools, manually translating visual ideas into parameters, and losing the context of an ongoing conversation. This project brings generation, editing, reference workflows, model comparison, and asset management into a DeepSeek Harness conversation and canvas.

Key features

• Chat-Based Image Creation and Editing — Describe an image, reference image, or requested edit in natural language. The agent supports text-to-image, image-to-image, multi-image references, style transfer, and iterative revision.

  Source: README

• Batch Generation and Prompt Inspiration — Create a planned set of images from one request, with a limit of 10 images per call. The built-in library contains 500+ reusable prompts that the agent can retrieve or the user can browse, favorite, and copy.

  Source: README

• Context-Aware AI Canvas — Draw a rough sketch, place reference images on an infinite canvas, or mark the region that should change. The agent can read the canvas selection and turn sketches into finished artwork or edit only the marked area.

  Source: README

• Studio, Gallery, and Model Comparison — Use a dedicated workspace for direct control over provider, model, ratio, quality, and watermark settings. Results can be searched, favorited, edited, regenerated, compared across models, and managed in a workspace-scoped gallery.

  Source: README

• Animation and Local Workflow Support — Supported OpenAI and compatible channels can produce looped GIF animations that can be played, downloaded, and revised with frame-specific comments. Named ComfyUI workflows can also be imported and selected directly from chat.

  Source: README

Quick start

Copied verbatim from the project's README.

pnpm dsh plugin --profile web add dsh-image-gen@latest

画一张雨夜霓虹街头的赛博朋克猫咪，电影感光线，16:9。

给我的小红书笔记《周末露营清单》配 5 张图：帐篷、炉具、睡袋、照明、零食，统一日系手绘风，3:4。

Who it may suit

This may suit DeepSeek Harness users who want to create or edit images through conversational workflows, as well as creators who need canvas-based iteration, prompt libraries, model comparison, or local ComfyUI integration.

Topics to discuss

• Which conversational image-generation workflows should receive priority as the plugin evolves?

• How should provider-specific capabilities be presented for comparison and discovery?

• Which privacy, download-security, and workspace-management defaults best fit self-hosted users?

Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.

Read the original README https://github.com/shanliuling/dsh-image-gen#readme

ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.