KohakuBlueleaf/z-tipo-extension

What it solves

It simplifies the process of creating detailed prompts for image generation models. Instead of manually writing long lists of tags or complex descriptions, users can provide a few basic tags or captions, and the tool automatically expands them into comprehensive prompts.

How it works

The extension integrates LLM-based systems—TIPO and DanTagGen—into popular image generation interfaces like SD-WebUI, Forge, and ComfyUI. TIPO handles both tags and natural language to optimize prompts, while DanTagGen specifically generates Danbooru-style tags. Users can control the output length, ban specific tags using regex, and customize the prompt format using special placeholders for categories like characters, artists, and quality markers.

Who it’s for

Digital artists and AI image creators who use Stable Diffusion-based tools and want to generate high-quality, detailed prompts without manual effort.

Highlights

• Supports multiple interfaces: SD-WebUI, Forge, and ComfyUI.
• Dual model support: Uses TIPO for flexible natural language/tag mixing and DanTagGen for specific tag generation.
• Customizable output: Adjustable tag and natural language length, seed control for sampling, and temperature settings for creativity.
• Flexible formatting: Uses a template system to organize prompts into specific categories like copyright, artist, and quality.