Architect's Toolkit: The Precision Wildcard System

"Stop fighting your model, start directing it."

The Architect's Toolkit is not just a collection of wildcards—it is a modular, high-precision framework designed for users who value creative control over random output.

Frustrated by "mushy" images, inconsistent results, and models ignoring your core prompts? This toolkit was built to solve the "prompt noise" problem. By stripping away redundant quality-fluff and technical ballast, it forces your model to focus on what actually matters: Composition, Scenario, and Intent.

Key Features:

Modular Architecture: Designed for logical, stacking prompts. Build your scene from the ground up, starting with core actions and scenarios rather than generic quality tags.
Zero-Noise Design: Optimized for modern SDXL models (like Juggify). We have purged standard "masterpiece" and "8k" tags that clutter your token limit and dilute your prompt's impact.
Action-Oriented Logic: Every wildcard is crafted to be an instruction, not just a description. This toolkit is optimized for users who want to define what happens in the frame.
Consistent Output: By utilizing strict categorization, you achieve a higher success rate per generation. Stop wasting your VRAM on variations of a blurry mess.

How to use it:

The Architect's Toolkit works best with a clean, minimalist prompting style. Move away from long-string prompt spaghetti.

Define the Scene: Use the Master Scenarios to set the atmosphere.
Define the Action: Use the Action-Objects for narrative depth.
Define the Composition: Use the Perspective views to frame your shot.

This toolkit is for the intentional artist. It transforms your Stable Diffusion workflow from a chaotic lottery into a reliable, modular creative engine.

Compatible with Stability Matrix / WebUI Forge.

Note: Please note that I cannot guarantee full functionality on other systems. It works flawlessly on my setup with an RTX 3060 (12GB) and an i5-12400F.

⚠️ Important Usage Warning: Keep it Focused!

Avoid chaining multiple "Master" categories in a single prompt. While the toolkit is modular, the model's attention span is finite. Trying to combine too many complex "Master" scenarios at once will lead to conflicting instructions, diluted details, and a loss of visual coherence. Pick one primary scenario/perspective per generation to maintain high quality and narrative clarity.

Example for a clean input: __MASTER_BDSM__, __MASTER_PERSPECTIVE_VIEW__, (dramatic interrogation scene:1.3), (cool blue edge light:1.2)

P.S.: --- Please didnt use Prompt Magic

Description

Architects Toolkit v3.0 – The Precision Engine

v3.0 is a complete architectural overhaul of the system. I have restructured the entire collection from the ground up to provide a logical, consistent foundation for your prompts. This is no longer just a loose collection of files – it is a highly curated composition engine consisting of over 30,000 professional prompt building blocks.

What's new in v3.0?

The 30k+ Library: Expanded to over 30,000 hand-curated prompt building blocks.
Logical Hierarchy: The entire folder structure has been standardized with an ID-based system (01-Subject, 02-Pose, etc.). This ensures your AI respects the logical flow of your composition.
Curated Content Split: * Items 01-10: Dedicated to the "Cyber-Goth" theme – providing a high level of consistency for this specific aesthetic.
- Items 11-20: A massive pool of randomly generated/diverse components, allowing for endless creative experimentation across various styles.
Plug & Play: Optimized for native use with sd-dynamic-prompts – simply drop it into your folder and start generating.
Bonus Automation: Included an optional FULL-Random.py script for advanced users who want to automate the logical prompt structure and copy it directly to their clipboard.
Juggify-Ready: Perfectly tuned for high-quality models like Juggernaut XL-based merges. Clean, high-fidelity results with minimal negative prompting.

How to use

Installation: Copy the MASTER_COLLECTION folder into your sd-dynamic-prompts/wildcards directory.
Implementation: Use the standard wildcard syntax (__MASTER_COLLECTION/filename__) in your prompt.
The "Architect" Method: Build your prompt blocks in a sequence (e.g., 01-Subject, 02-Action_Pose, 05-Environment). By following this numerical order, you guarantee that the composition remains stable while the individual elements vary.

Und hier ist die passende README.txt, die du in deine ZIP-Datei packst:

Architects Toolkit v3.0 - The Precision Engine

Thank you for downloading the Architects Toolkit v3.0. This collection is a highly curated composition engine designed to give you maximum control over your image generation process.

Overview

This toolkit contains over 30,000 hand-picked prompt building blocks. They are structured using a unique ID-based system (01-Subject, 02-Pose, etc.) to ensure that your image composition remains stable, logical, and consistent across thousands of variations.

How to use

Installation: Copy the MASTER_COLLECTION folder into your sd-dynamic-prompts/wildcards directory.
Implementation: Use the standard wildcard syntax (__MASTER_COLLECTION/filename__) in your prompt.
The "Architect" Method: For the best results, structure your prompt sequentially (e.g., 01-Subject, 02-Action_Pose, 05-Environment). By following the numerical order, you guarantee that the AI builds your image logically.

Optional: Automation Script (Advanced)

Included in this download is a FULL-Random.py script.

What it does: It automatically generates a professionally structured prompt, ensures the logical sequence (01-XX) is maintained, and copies the result directly to your clipboard.
Requirement: You need Python installed on your system.
Usage: Run the script and paste the output directly into your WebUI prompt field.

Tips for success

Negatives: This toolkit is designed to be "clean." You only need to add negative prompts for common artifacts (like text or watermarks).
Recommended Settings: For best results with high-quality models (e.g., Juggernaut XL-based merges), use:
- Steps: 28
- Sampler: DPM++ 3M SDE Karras
- Resolution: 1344x768

Enjoy your generative journey and build your own worlds!