Model
Trends
.ai
Model
Trends
.ai
the best open source ai models

ReActor Fast Face Swap node in ComfyUI

ReActor swaps a face from a source photo onto the generated picture with InsightFace, with optional restoration (GFPGAN, CodeFormer) and face boosting, for consistent characters without training a LoRA.
Node pack: ComfyUI-ReActor
Category: Popular custom node packs

What ReActor Fast Face Swap does

ReActor (ComfyUI-ReActor, previously comfyui-reactor-node) runs the inswapper_128 model: it detects faces in the input image and the source image, swaps the identity, then runs a restoration model to clean up the low-resolution swap. Options select which face index to swap, gender detection, restore_face model and visibility, and codeformer_weight. A Face Boost option re-upscales the swapped face.
Use it after your sampler and before Save Image: generate the scene with any face, then swap in the character. It is fast (a second per face) and works with any base model since it operates on pixels. For the best quality combine it with a FaceDetailer pass at low denoise after the swap.

Inputs

Name
Type
What it is
input_image
IMAGE
The generated picture.
source_image
IMAGE
A photo of the face to transplant (or a saved face model).
swap_model
COMBO
inswapper_128.onnx.
facedetection
COMBO
retinaface_resnet50 and others.
face_restore_model / face_restore_visibility / codeformer_weight
COMBO / FLOAT
Post-restoration settings.
detect_gender_input / detect_gender_source / input_faces_index / source_faces_index
COMBO / STRING
Which faces to swap, 0 is the largest or leftmost.

Outputs

Name
Type
What it is
IMAGE
IMAGE
The swapped picture.
FACE_MODEL
FACE_MODEL
Reusable embedding of the source face.
AD
Skip the setup: ComfyUI workflows preinstalled on the cloud
BitVector runs ready-made ComfyUI workflows on its own GPUs. No install, no missing nodes, no red boxes. Open it on your phone or laptop and generate in a minute.

How to use ReActor Fast Face Swap

  1. Install ComfyUI-ReActor from Manager; on Windows install the insightface wheel and Visual C++ build tools if prompted.
  2. Add ReActor Fast Face Swap after VAE Decode; connect input_image and a source_image from Load Image.
  3. Pick inswapper_128.onnx and restore with GFPGANv1.4 at visibility 1, codeformer_weight 0.5.
  4. Queue; adjust input_faces_index when the wrong person is swapped.
  5. Optional: follow with FaceDetailer at denoise 0.3 for sharper skin.

Settings and tips

  • A frontal, well-lit source photo gives the best swap.
  • Save a FACE_MODEL with Save Face Model and reuse it across workflows.
  • The swap is 128 px: the restoration step matters; try CodeFormer for sharpness, GFPGAN for naturalness.
  • For a group picture, set input_faces_index to "0,1" and source_faces_index to "0,0".

Troubleshooting ReActor Fast Face Swap

ModuleNotFoundError: insightface or onnxruntime

Why it happens
The Python dependencies did not compile; common on Windows without build tools.

How to fix it
Install the prebuilt insightface wheel for your Python version into the ComfyUI environment and onnxruntime-gpu, then restart.

No face detected in source image

Why it happens
The photo is too small, side-on or occluded.

How to fix it
Use a larger frontal photo; crop it to the face area.

The swapped face is blurry or plastic

Why it happens
The inswapper output is 128 px and the restorer smooths it.

How to fix it
Enable Face Boost or follow with FaceDetailer at denoise 0.3; lower face_restore_visibility to 0.7.

The node is red / NSFW filter error

Why it happens
The pack has an NSFW detector step that needs its model download; failed downloads block the node.

How to fix it
Check the console for the download URL, place the model where indicated and restart.

AD
Run this workflow from your phone
Every workflow on this page is preinstalled on BitVector with the models and custom nodes already in place. Pick one, type a prompt, done. Zero setup, nothing to download.

Questions about ReActor Fast Face Swap

Is ReActor better than a LoRA?

It is faster and needs one photo, but copies only the face region at 128 px. A LoRA captures body, hair and expression and renders at full quality.

Can it swap faces in video?

Yes, per frame: batch the frames from Load Video and swap, then recombine with Video Combine.

Related nodes

FaceDetailer (Impact Pack)
ComfyUI-Impact-Pack
FaceDetailer finds faces (or hands, eyes) with a YOLO detector, crops each one, re-renders it at full resolution with your model and prompt, and pastes it back, fixing the small, mushy faces in wide shots.
Load Image
ComfyUI core
Load Image reads a picture from the ComfyUI input folder (or an upload) and outputs an IMAGE tensor plus a MASK taken from the alpha channel, the starting point for img2img, inpainting, ControlNet and IP-Adapter graphs.
Save Image
ComfyUI core
Save Image writes the IMAGE tensor to ComfyUI/output as a PNG, with the whole workflow embedded in the file metadata so the picture can be dragged back into ComfyUI to restore the graph.
Video Combine (VideoHelperSuite)
ComfyUI-VideoHelperSuite
Video Combine turns an IMAGE batch of frames into an MP4, WebM, GIF or image sequence at a chosen frame rate, with optional audio, loop count and ping-pong, the standard output node for every video workflow.
AD
Tired of fixing nodes? Let the cloud do it
BitVector keeps hundreds of ComfyUI workflows installed, updated and tested on fast cloud GPUs. No Python, no CUDA errors, no VRAM limits. Works from any browser.

More ComfyUI nodes

CLIP Text Encode (Prompt)
ComfyUI core
CLIP Text Encode turns a text prompt into CONDITIONING using the model text encoder. One node holds the positive prompt, a second one the negative prompt.
ComfyUI Manager
ComfyUI-Manager
ComfyUI Manager is the extension that installs, updates and fixes custom node packs and models from inside the interface, resolves missing nodes in imported workflows and snapshots your setup.
KSampler
ComfyUI core
KSampler runs the denoising loop: it takes the model, prompts and a latent and produces the finished latent image, controlled by seed, steps, cfg, sampler, scheduler and denoise.
Load Checkpoint
ComfyUI core
Load Checkpoint (CheckpointLoaderSimple) opens a .safetensors or .ckpt model file and hands out the three parts every workflow needs: the diffusion MODEL, the CLIP text encoder and the VAE.
VAE Decode
ComfyUI core
VAE Decode converts the sampled LATENT into a pixel IMAGE with the VAE; VAE Decode (Tiled) does the same in tiles for very large images.
Empty Latent Image
ComfyUI core
Empty Latent Image creates the blank latent canvas (width, height, batch_size) that text-to-image sampling starts from; dimensions must be multiples of 8 and match the model family.