Skip to content
WorkflowListedExternal
Open source

Video to Video with Face Swap in ComfyUI — v1.0

Showing how to do video to video in comfyui and keeping a consistent face at the end.

Sourcecivitai.com/models/161796
TaskMultiCapabilitiesUpscaling · Video to videoVersionRelease 182135 — Video to Video Workflow.jsonEstimated VRAMNot yet characterizedDgrade

Runtime profile

Access
External workflow
Status
Listed
Version
Video to Video Workflow.json
Source
Civitai
Updated
Oct 18, 2026

Source description

Showing how to do video to video in comfyui and keeping a consistent face at the end.

We keep the motion of the original video by using controlnet depth and open pose.

We use animatediff to keep the animation stable.

Finally ReActor and face upscaler to keep the face that we want. Optionally we also apply IPAdaptor during the generation to help keep the face closer to what we want even before the swap. (Note: IPAdaptor can greatly slow down the generation depending on your machine)

What you need

Model files
5.2 GiB known subtotal

Remaining hardware or download-size estimates are not available yet.

Requirements

Models

8

  • control_depth-fp16.safetensors

    Possible

    ControlNet · Unknown

  • control_openpose-fp16.safetensors

    Possible

    ControlNet · Unknown

  • GFPGANv1.4.pth

    Not resolved

    Checkpoint · Unknown

  • inswapper_128.onnx

    Found

    Checkpoint · 529 MB · ONNX · fp32 · Unknown

  • ip-adapter-plus-face_sd15.bin

    File unverified

    LoRA · Unknown

  • juggernaut_final.safetensors

    Possible

    Checkpoint · 1.99 GB · Civitai · 46422

  • mm_sd_v15_v2.ckpt

    Found

    Checkpoint · 1.69 GB · CKPT · fp32 · full · Unknown

  • model.safetensors

    Possible

    Vision model · Unknown

AI-generated commentary

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

Purpose

  • Use this ComfyUI workflow to turn an original video and a desired face into a video-to-video face swap that carries the original motion into the resulting video.

Flow

  • ControlNet depth and open pose carry motion from the original video into the result.

  • The workflow pairs AnimateDiff for animation stability with ReActor and a face upscaler for keeping the chosen face.

Inputs and outputs

  • The original video supplies the motion; the intended output is a video with the chosen face retained.

Setup

  • Find version v1.0 by A1lazydog on Civitai: https://civitai.com/models/161796?modelVersionId=182135.

Dependencies

  • The workflow requires these listed model files: model.safetensors, mm_sd_v15_v2.ckpt, and inswapper_128.onnx.

  • Other model filenames listed for the workflow are control_depth-fp16.safetensors, control_openpose-fp16.safetensors, GFPGANv1.4.pth, ip-adapter-plus-face_sd15.bin, and juggernaut_final.safetensors.

Configuration

  • IPAdapter is optional; during generation it can help keep the generated face closer to the face you want before the swap.

Cautions

  • IPAdapter can greatly slow generation, depending on your machine.

Does this need to be edited?

Sign in to send an edit request.

Sources

1 source
CivitaiPrimaryVersion 182135
civitai.com/models/161796

Verified Oct 18

Source excerpts

2 excerpts
NoteOct 18civitai.com/models/161796

Source context: 8165 downloads · Type Workflows · Base model Other

NoteAug 9civitai.com/models/161796

Source context: 7892 downloads · Type Workflows · Base model Other