WorkflowListedExternal

AITK-Flux Faceswap

a workflow with flux redux, fill and the new ace++ lora in a single workflow

Sourcecivitai.com/models/1250334

TaskMultiCapabilitiesImage edit · Image to image · UpscalingFamilyFLUXEstimated VRAMEstimate unavailableBgrade

Runtime profile

Access: External workflow
Status: Listed
Source: Civitai
Updated: Jul 20, 2026

Source excerpt

a workflow with flux redux, fill and the new ace++ lora in a single workflow

Estimated VRAM requirement

Estimate unavailable

9.12 GB across 1 of 13 model files. Model file total + 25% loading overhead + 2 GB execution buffer, rounded up.

Requirements

56 requirements

Models

AI-generated commentary

AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.

Purpose

AITK-Flux Faceswap is a workflow for face swapping and image editing. It combines flux redux, fill, and the new ace++ lora, takes a loaded image through LoadImage, and names PreviewImage as a preview output.

Flow

The graph includes image loading and size handling, face analysis and location, face-embedding distance, segmentation, mask creation, cropping, inpainting, sampling, compositing, decoding, and preview stages.
It also names CLIP text and vision encoders, FLUX guidance, style-model loading and application, image resizing, megapixel scaling, and an upscaler loader.

Inputs and outputs

The named image-handling stages include LoadImage for image input, PreviewImage for preview output, MaskPreview+ for mask preview, and Image Comparer (rgthree) for comparison.

Setup

Find AITK-Flux Faceswap on Civitai at https://civitai.com/models/1250334.
The listed publisher is Archi Teknik101, and the base model is Flux.1 D.

Dependencies

The workflow lists these model files: 4x-ClearRealityV1.pth; ae.safetensors; flux1-fill-dev (1).safetensors; flux1-redux-dev (1).safetensors; pixelwave_flux1Dev03.safetensors; sigclip_vision_patch14_384 (1).safetensors; t5xxl_fp16.safetensors; ViT-L-14-TEXT-detail-improved-hiT-Gm
Named node components include BasicGuider, BasicScheduler, CFGGuider, CLIPTextEncode, CLIPVisionEncode, CLIPVisionLoader, ConditioningZeroOut, DualCLIPLoader, ExpressionEditor, FaceAnalysisModels, FaceBoundingBox, FaceEmbedDistance, FaceSegmentation, Fast Groups Bypasser (rgthree
Other named node components include ImageResize+, ImageResizeKJ, ImageScaleToMegapixels, ImageToMask, ImpactSwitch, InpaintModelConditioning, JWInteger, KSamplerSelect, LoadImage, MaskPreview+, Masks Add, MaskToImage, Power Lora Loader (rgthree), PreviewImage, RandomNoise, Samper

Configuration

Review FluxGuidance, SamplerCustomAdvanced, ImageResize+, and Power Lora Loader (rgthree) when adjusting guidance, sampling, image resizing, or LoRA loading.
Suggestion · not verified
Review InpaintModelConditioning, UpscaleModelLoader, FaceSegmentation, and StyleModelAdvancedApply when adjusting inpainting, upscaling, face segmentation, or style-model application.
Suggestion · not verified

Cautions

Before using it, verify that JWInteger, Image Comparer (rgthree), GrowMaskWithBlur, and StyleModelLoader are available in your setup.
Suggestion · not verified
Verify the exact file and type you intend to use for 4x-ClearRealityV1.pth, listed as an upscaler, and ViT-L-14-TEXT-detail-improved-hiT-GmP-HF.safetensors, listed as a text encoder.
Suggestion · not verified

Does this need to be edited?

Sources

1 source

CivitaiSource

civitai.com/models/1250334

Verified Jul 20

Source excerpts

1 excerpt

NoteJul 20civitai.com/models/1250334

Source context: 858 downloads · Type Workflows · Base model Flux.1 D