🛠️ Desaign’s Z-Image Turbo Workflow — ZIT_alpha Release

The Desaign ZIT_alpha release introduces Desaign’s newest text-to-image and image-to-image engine, built around a highly optimized sampler/scheduler configuration that delivers exceptional stability, coherence, and detail across all LoRAs, regardless of subject or style. This configuration is the result of extensive testing, and we consider it to be the current optimal baseline for ZIT.

this workflow is designed to be fast, clean, coherent, and reliable, while remaining extremely simple to operate. It is ideally suited for everything from explorative concept sketching to polished cinematic rendering.

🔄 Overview

Desaign ZIT_alpha is organized around two rendering passes:

1. First Pass — TXT2IMG or IMG2IMG

A flexible entry point that lets you either:

• generate from text directly, or

• transform an existing image with strong structural preservation.

2. Second Pass — Refinement

A controlled re-render that reinforces coherence, texture density, and global aesthetic quality.

Between these two passes sits a lightweight ×1.5 upscale stage, which improves detail continuity and stabilizes features without increasing render time significantly.

The workflow uses subgraphs, toggles, and sliders to keep the interface clean and intuitive, allowing both beginners and power users to navigate it quickly.

---

⚙️ Technical Highlights

✓ Optimal ZIT Sampler/Scheduler Configuration

ZIT_alpha uses a tuned combination of sampler and scheduler parameters that consistently outperform typical configurations in clarity, texture retention, and structural stability.
These settings are LoRA-agnostic and work out of the box with any model extension.

✓ Modular Pass Architecture

Each pass is optional and controlled by simple toggles.
You may use only the first pass for quick sketches or enable the second pass for cinematic refinement.

✓ Upscale-Before-Refine Workflow

A ×1.5 upscale between passes will enhances micro-detail and coherence when activating the second pass.
This has become the recommended baseline for ZIT.

✓ IMG2IMG Denoise Tolerance

ZIT is unusually permissive with denoise values:

• You can push IMG2IMG denoise up to 0.75 without losing global structure.

✓ High-Creativity Second Pass

The refinement stage is calibrated for expressive re-interpretation:

• Recommended denoise range: 0.6–0.8

• Sweet spot: 0.7 for texture richness, better “baked-in” detail, and creativity.

✓ Interface Convenience

• Mode selectors

• Sliders for denoise, upscale value

• Optional toggles for enabling/disabling modules

• Subgraphs for clarity and reduced clutter