Run this workflow on InstaSD
Get started in minutes! Run this ComfyUI workflow online - no setup required.
Description
Effortlessly transfer subjects from one image to another with ComfyUI ACE++ & Redux. This workflow uses advanced extraction and blending techniques to preserve key details—like pose and composition—so you can seamlessly import subjects into entirely new scenes or styles. No more tedious manual masking!
🎯 Features
- Subject Extraction Isolate people, objects, or characters from a source image with minimal artifacts.
- Flexible Blending Adapt subjects to new backgrounds, styles, or lighting conditions while maintaining realism.
- Parameter Control Fine‑tune blending strength, color matching, and edge detailing through ComfyUI’s node interface.
- Style Agnostic Works with a variety of art styles—photorealistic, anime, illustration, and more.
💡 Use Cases
- Creative Collages Combine elements from different artworks or photographs into original compositions.
- Concept Art Quickly populate scenes with existing characters for storyboards or project mockups.
- Photo Restoration Migrate restored subjects into new images for family portraits, memorials, or creative tributes.
⚙️ How It Works
- Upload Source & Target Provide an image containing the subject(s) to migrate and a target image where you’d like them placed.
- Configure ACE++ & Redux Adjust extraction parameters for edge detection, color matching, and intensity.
- Blend & Refine Use ComfyUI’s node-based interface to refine subject positioning, scale, and style matching.
- Export or Further Edit Finalize your composite image or export for post-processing in external editors.
Credits : Horror_Dirt6176
Models
| File | Destination | Source |
|---|---|---|
| ae.safetensors | /ComfyUI/models/vae | Download |
| flux1-fill-dev.safetensors | /ComfyUI/models/unet | Download |
| sigclip_vision_patch14_384.safetensors | /ComfyUI/models/clip_vision | Download |
| flux1-redux-dev.safetensors | /ComfyUI/models/style_models | Download |
| BiRefNet-general-epoch_244.pth | /ComfyUI/models/loras | Download |
| comfyui_subject_lora16.safetensors | /ComfyUI/models/loras/subject | Download |
| clip_l.safetensors | /ComfyUI/models/clip | Download |
| t5xxl_fp8_e4m3fn.safetensors | /ComfyUI/models/clip | Download |
| BiRefNet-general-epoch_244.pth | /ComfyUI/models/BiRefNet | Download |
| BiRefNet-general-epoch_244.pth | /ComfyUI/models/BiRefNet/pth | Download |
Nodes
ImageResizeKJGrowMaskWithBlurMaskToImageImageConcanateImageToMaskEmptyImageLoadImageLayerMask: BiRefNetUltraV2LayerMask: LoadBiRefNetModelGetImageSize+SolidMaskImage OverlaySaveImageKSamplerVAEDecodeImageCrop+DualCLIPLoaderUNETLoaderCLIPVisionLoaderStyleModelLoaderLoadLoraFromHFCLIPTextEncodeConditioningZeroOutFluxGuidanceInpaintModelConditioningVAELoaderDifferentialDiffusionCLIPVisionEncodeStyleModelApply