shanliuling/dsh-image-gen

AI image studio for DeepSeek Harness — generate, edit & compare images in chat, with 500+ prompts, gallery, multi-model workflows and ComfyUI.

What it solves

dsh-image-gen is a comprehensive AI image creation suite designed as a plugin for DeepSeek Harness (DSH). It transforms DSH from a simple chat interface into a full-fledged image production workflow, enabling users to move from natural language descriptions to precise parameter control, batch generation, and organized asset management.

How it works

The plugin integrates multiple AI image generation providers via a BYOK (Bring Your Own Key) model. It offers four primary interaction modes:

  • Chat-based Generation: Users can generate images, perform image-to-image tasks, and conduct continuous editing using natural language through an Agent.
  • Studio Workbench: A dedicated space for precise parameter control, batch generation, and the use of multiple reference images.
  • Inspiration Library: A curated collection of over 500 prompt examples that users can browse and import directly into the Studio.
  • Gallery Management: A centralized system to search, filter, and organize generated images, with support for workspace isolation.

It supports cloud providers like Google Gemini, OpenAI, ByteDance Seedream, and Aliyun DashScope, as well as local ComfyUI workflows via API Format Workflow JSON.

Who it’s for

Users of DeepSeek Harness who want to integrate professional AI image generation and editing workflows directly into their AI agent environment, ranging from casual creators to those using local GPU power via ComfyUI.

Highlights

  • Continuous Editing: Modify prompts to regenerate images and switch between historical versions within the same card.
  • Multi-Model Comparison: Run the same prompt and reference images across multiple models simultaneously to compare results on one canvas.
  • Batch Production: Generate multiple candidates in the Studio and save only the desired results.
  • Local ComfyUI Integration: Support for multiple named workflows allowing agents to select specific workflows for text-to-image or image-to-image tasks.
  • Prompt Library: 500+ built-in examples for style and composition inspiration without consuming tokens during browsing.

Related

  • Project
    yanokusnir-ai/one-node-flux-2-klein

    A ComfyUI custom node that wraps the full FLUX.2 [klein] workflow into a single, self-contained UI widget to eliminate complex node graphs.

  • Project
    wooyeolbaek/attention-map-diffusers

    attention‑map‑diffusers is a Python library that captures and visualises the attention tensors of Hugging Face Diffusers image and video generation pipelines, letting users see how prompts, image patches, or video frames influence each other.

  • Project
    qurator-spk/eynollah

    Eynollah is a Python library and command‑line tool for deep‑learning based document layout analysis, image binarization, enhancement, OCR, and reading‑order detection. It supports 10 layout classes, offers both CNN‑RNN and Transformer OCR models, outputs PAGE‑XML, and can be installed via PyPI or Docker. Pre‑trained ONNX models enable fast inference; full TensorFlow/PyTorch checkpoints are available for fine‑tuning. The project targets high‑quality processing of historical and heterogeneous scanned documents.

  • Project
    Extraltodeus/ComfyUI-AutomaticCFG

    A ComfyUI extension that prevents image artifacts and over-saturation by automatically rescaling CFG and provides tools to accelerate generation speed.