NMKD

NMKD Stable Diffusion GUI is a free interface for running Stable Diffusion locally on your own hardware. Dependencies are bundled, so installation avoids the usual dependency maze, and the tool supports text-to-image, image-to-image, and instruction-based editing. Prompt handling includes emphasis and negative prompting, and you can load custom Stable Diffusion models, custom VAE models, LoRA concepts and characters, and text inversion concepts. Built-in RealESRGAN upscaling plus face repair with CodeFormer or GFPGAN round out the feature list, alongside a prompt queue and history. The listing states no data is shared or collected, so generation stays on your machine.

- Generate images from text or edit existing images

- Apply instruction-based edits through prompting

- Load custom models, VAE files, LoRA, and concepts

- Upscale and repair faces without a second tool

- Run everything offline on your own GPU

- People with a capable GPU who want local generation

- Users who want their prompts and images to stay private

- Tinkerers who load custom models and LoRA regularly

The listing is explicit that output is completely uncensored and unfiltered and that the author takes no responsibility for generated content. That responsibility is yours, and it includes not creating imagery of real people without consent, nothing sexual involving minors, and nothing that breaks the law or the rules of wherever you publish. You need a suitable GPU, and generation speed scales with it, so an older card means waiting. The project is under active development, so expect minor issues and occasional changes, and follow the system requirements guide before installing.