This extension for AUTOMATIC1111's Stable Diffusion WebUI provides a powerful template language for generating diverse and combinatorial prompts. It targets users who want to explore a wide range of creative possibilities by automating prompt variation, enabling efficient experimentation with styles, artists, and concepts.

How It Works

The core of the extension is a template language that supports random selection from lists (e.g., {option1|option2}), recursive wildcard file matching (e.g., __colors*__), and combinatorial generation of all possible prompt combinations. This allows users to define complex prompt structures that can be expanded into numerous unique prompts, facilitating systematic exploration of the latent space.

Quick Start & Requirements

• Installation: Install via the WebUI's Extensions tab or manually clone the repository into the extensions/sd-dynamic-prompts directory.
• Prerequisites: AUTOMATIC1111/stable-diffusion-webui. Optional dependencies for specific features like Magic Prompt or Attention Grabber can be installed via pip install dynamicprompts[attentiongrabber,magicprompt].
• Resources: Wildcard files can be managed through the "Wildcards Manager" tab. Bundled collections like "jumbo," "parrotzone," and "devilkkw" are available.
• Documentation: More complete documentation and tutorials are available.

Highlighted Details

• Supports fuzzy globbing and recursive wildcard matching for flexible file selection.
• Enables combinatorial generation to produce every possible prompt permutation.
• Integrates with other WebUI features like X/Y Plot for advanced parameter testing.
• Offers optional prompt expansion models (e.g., MagicPrompt) for AI-assisted prompt generation.

Maintenance & Community

The project is actively maintained. Community resources include a GitHub repository for issues and discussions.

Licensing & Compatibility

The extension is released under the MIT License, permitting commercial use and integration with closed-source projects.

Limitations & Caveats

Relative paths for wildcard files are not currently supported; full paths are required. Some AI prompt generation models may require significant VRAM. Combinatorial generation with variation strength > 0 might require specific configuration adjustments for desired results.