🧠 ABSOLUTE SYNTHETIC INTUITION (ASI)
Modular Prompt System for Z Image Turbo
What is ASI?
ABSOLUTE SYNTHETIC INTUITION (ASI) is a modular prompt architecture designed to structure image generation through logic, weighting, and separation, instead of relying on monolithic prompts.
ASI is a system of wildcards, not a single prompt.
It is optimized for Z Image Turbo and focuses on:
stability
efficiency
reproducibility
clean subject dominance
🚀 Quick Start (Important)
If you only want to use ASI, you only need one file:
✅ Master Wildcard (Entry Point)
__ASI-Master-Orchestrator__
or the Turbo-optimized version:
__ASI-Master-Orchestrator-v2__
This master wildcard automatically assembles:
style handling
logic modules
render descriptors
weight logic
optional negatives
All other files are internal modules used by the orchestrator.
🎯 ASI Design Principles
ASI is built on three core principles:
1. Modularity
Each function is isolated into its own wildcard:
style
logic
render
isolation
depth
No monolithic prompt blocks.
2. Dichotomy Weight Logic
ASI uses a strict separation between:
high-impact weights
low-impact weights
There are no ambiguous mid-values.
This ensures predictable behavior across generations.
3. Model Awareness
ASI is explicitly designed around the behavior of Z Image Turbo, including:
token sensitivity
style bleed tendencies
depth compression
🧩 System Structure (Overview)
Shared Core
__ASI-Weight-High____ASI-Weight-Low__Style libraries (Front / Back)
These elements are reused across all ASI pipelines.
Generation 1 – Classic Modular
Fully modular
Multiple render stacks
Designed for flexibility and experimentation
Generation 2 – Turbo Optimized (Recommended)
Reduced token count
High-impact clustering
Background style weighted at 0.7
Strong subject dominance and 2.5D separation
🔍 Core ASI Logic Modules
ASI Logic – Isolation
Prevents:
color bleeding
style contamination
flat subject/background merging
Ensures clean layer separation.
ASI Logic – PopOut
Creates:
perceived depth
subtle relief
foreground dominance
cinematic subject presence
Especially effective for character-focused images.
🎨 Style Handling (ASI Concept)
Style-Front defines the primary visual language
Style-Back reinforces coherence at reduced strength
This avoids overpowering the subject while maintaining consistency.
❌ Why There Is No Mandatory Negative Wildcard
ASI intentionally avoids large negative prompt blocks.
Reasons:
Z Image Turbo responds better to positive structural guidance
Most common failure modes are prevented by:
isolation logic
weight separation
controlled style reinforcement
Minimal negatives can still be added manually if desired.
📂 Index File (Documentation Only)
The ASI index file is:
a human-readable overview
not part of the generation pipeline
included purely for orientation and reconstruction
It has no functional impact.
🧪 Intended Use Cases
Character design
Concept art
Cinematic portraits
Modular style experimentation
Fast iteration workflows
🏁 Final Notes
ASI is designed to be:
extended
modified
adapted
It is a framework, not a closed system.
If you build on it, that is expected behavior.
💬 Feedback & Iteration
If you create:
custom modules
alternative pipelines
optimizations for other models
feel free to share them.
ASI evolves through use.
Pro-Tip: Workflow Optimization & Roadmap
Pro-Tip for Maximum Efficiency: Let the AI handle the structural heavy lifting. When using an LLM to write your prompts, simply provide the full list of available wildcards. The AI will automatically distribute and position them to achieve the best possible balance based on your specific focus. This automated orchestration ensures the most stable and high-quality results within the ASI framework.
Background & Roadmap: This system is the direct evolution of The Loom - Quality, Style & Perspective and is currently my top recommendation for prompt architecture.
Status: Currently in Beta (fully functional and highly stable).
Release: The final integration is scheduled for late January as a major update to the main module: The Loom – Modular Prompt Engine (All-in-One Edition).
Future Updates: The final version will feature significant refinements, specifically focusing on advanced glow effects and enhanced light-logic.
Pro-Tip: Style Handling & LoRA Compatibility
Multi-Style Synergies: The wildcards are designed to combine multiple visual styles seamlessly. This architecture works best in its standard version, where the internal logic maintains perfect balance.
LoRA Interaction: Note that using external LoRAs will generally not cause significant style shifts or drastic changes to the aesthetic. The ASI framework is built for high structural integrity, making it resistant to style bleeding.
Exception: High-impact "Punk" styles (e.g., Cyberpunk, Steampunk) are the only categories where LoRAs will exert a dominant influence on the final output.