Extract subjects from existing RGB images directly into **true 4-channel RGBA PNGs with native alpha transparency** using **FLUX 3 Image Klein (5.6B)**!Unlike traditional segmentation masks <code>rembg</code>, <code>BiRefNet</code>) that only slice an existing image with a 2D mask, this LoRA leverages FLUX 3's native <code>ReferenceLatent</code> <code>image_cond</code>) conditioning together with a **4-channel RGBA VAE <code>flux2-klein-alpha-vae</code>)** to reconstruct clean, uncontaminated foreground edges, hair strands, glass, and semi-transparent details directly in latent space.---### ✨ Key Features- **Native 4-Channel RGBA Extraction:** Generates premultiplied RGBA latents decoded directly into 4-channel <code>RGBA</code> images <code>[B, H, W, 4]</code>).- **Prompt-Guided or Automatic Matting:** Extract the main subject automatically or specify via text prompt which exact object/character in the scene you want isolated.- **Clean Edges &amp; Semi-Transparency:** Eliminates background color spill/halos on hair, fur, glowing effects, and translucent surfaces.- **Native ComfyUI Integration:** Works seamlessly with ComfyUI's built-in <code>ReferenceLatent</code> node and <code>DenRakEiw_Nodes</code>](<a target="_blank" rel="ugc" href="https://github.com/DenRakEiw/DenRakEiw_Nodes">https://github.com/DenRakEiw/DenRakEiw_Nodes</a>).---### ⚙️ Recommended Settings- **Base Model:** <code>flux-3-image-klein.safetensors</code> (FLUX 3 Image Klein 5.6B)- **Text Encoder:** <code>qwen_3_5_4b.safetensors</code> (CLIPLoader type: <code>flux3</code>)- **LoRA File:** <code>flux3_klein_extract_lora.safetensors</code>- **LoRA Strength:** <code>1.0</code>- **Conditioning Setup:** Encode your input image with <code>Flux 2/3 RGBA VAE Encode</code> and attach it to the positive conditioning via ComfyUI's <code>ReferenceLatent</code> node.- **Sampler / Scheduler:** <code>euler</code> / <code>simple</code>- **Steps:** <code>20</code>- **CFG:** <code>3.5</code> (via <code>FluxGuidance</code>)- **Resolution:** <code>1024x1024</code> (or matching your input aspect ratio)---### 💬 Prompting Guide**1. Automatic Main Subject Extraction (Generic):**&gt; <code>Extract the main foreground subject, RGBA transparent background, isolated foreground object</code>**2. Targeted Subject Extraction (Specific):**&gt; <code>Extract foreground object, RGBA transparent background: a cute anime girl with silver hair and headphones</code>&gt; <code>Extract foreground object, RGBA transparent background: a glowing crystal potion bottle with golden Cork</code>---### 🧩 Required Custom Nodes &amp; RGBA VAE1. **Custom Nodes ([DenRakEiw_Nodes](<a target="_blank" rel="ugc" href="https://github.com/DenRakEiw/DenRakEiw_Nodes)):**">https://github.com/DenRakEiw/DenRakEiw_Nodes)):**</a>- <code>Load Flux 2/3 RGBA VAE (4-Ch Alpha)</code>- <code>Flux 2/3 RGBA VAE Encode (4-Ch Alpha)</code>- <code>Flux 2/3 RGBA VAE Decode (4-Ch Alpha)</code>- <code>Save Transparent PNG (RGBA)</code>- Ready-to-run workflow included in the repo: <code>workflows/Flux3_Klein_Extract_RGBA.json</code>2. **4-Channel RGBA VAE <code>flux2_klein_alpha_vae.safetensors</code>):**- Place in <code>ComfyUI/models/vae/</code>- Available on Hugging Face: [DenRakEiw/FLUX3-Image-Klein-Native-RGBA-Transparent-LoRA](<a target="_blank" rel="ugc" href="https://huggingface.co/DenRakEiw/FLUX3-Image-Klein-Native-RGBA-Transparent-LoRA">https://huggingface.co/DenRakEiw/FLUX3-Image-Klein-Native-RGBA-Transparent-LoRA</a>) or [trmz/flux2-klein-alpha](<a target="_blank" rel="ugc" href="https://huggingface.co/trmz/flux2-klein-alpha">https://huggingface.co/trmz/flux2-klein-alpha</a>)---### 🔗 Related Models &amp; Credits- **Text-to-RGBA Generation LoRA:** Check out the companion **FLUX 3 Image Klein Native RGBA Text-to-Transparent LoRA** for generating brand-new transparent assets from scratch!- **Credits:** Inspired by the research [Teaching a FLUX.2 VAE to See Transparency](<a target="_blank" rel="ugc" href="https://tr-mz.github.io/papers/flux2-klein-alpha/)*">https://tr-mz.github.io/papers/flux2-klein-alpha/)</a> by <code>trmz</code>. Adapted and trained for **FLUX 3 Image Klein (5.6B)** by **DenRakEiw**.

Files with SHA256: 4ba5a7f6c68fa57ae9ff9e5158da630dd1d2cb6f0c01f0cfdc3e16d089f16368