This workflow was created based on a request from the YouTube comments, asking for a low-VRAM setup with inpainting support, similar to my earlier low-VRAM workflows.

Overview

• GGUF-based workflow designed to run on low VRAM systems

• Supports Q4 and Q6 models

• Uses FP8 mixed text encoder for efficiency

• Can also run FP16 models if you have higher VRAM available

Workflow Design

• Built using only core, native nodes

• No missing nodes or external dependencies

• Loads and runs as-is

Inpainting Examples

• Adding sunglasses by masking the face area

• Changing clothing by masking the torso and prompting a new outfit

• Lighting and overall consistency are preserved across edits

Why This Workflow Exists

• Simple to read and modify

• Focused on practical image editing

• Optimized for systems with limited resources