ワークフローに戻る

Face Swap

The best instant AI face swap technology for photos.

ワークフローをダウンロードCloudで試す

Input

The base photo is resized to this many megapixels; aspect ratio is kept.

Advanced settings

On: Qwen 3.5 describes the base photo's facial expression and adds it to the swap prompt. Off: the plain head-swap instruction only.

Leave empty for a new result each run. Enter a number to repeat one.

Run

Output
Output
Example

例を試す

このワークフローについて

This ComfyUI workflow performs fast, high-quality face swaps on photos using a minimal, easy-to-tune graph. Two LoadImage nodes bring in your source face (the identity you want to transfer) and the target photo. A custom face-swap processor (node ID: 217f3726-59bc-4170-9ab1-c622bb578338) detects and aligns faces in the target, transfers the source identity onto the target while preserving expression, pose, and lighting, and outputs a composited image. An ImageCompare node lets you visually verify results against the original for quick A/B checks, and SaveImageAdvanced handles clean, repeatable exports.

Technically, the face-swap node encapsulates the core pipeline—face detection, alignment/warping, identity transfer, and blending—so you don’t need to manage separate detection or masking nodes. This makes the workflow compact and robust: you load images, optionally tweak the swap node’s settings (such as which detected face to replace or blend strength, depending on the implementation), preview with ImageCompare, and then save your output with consistent naming and format via SaveImageAdvanced.

It’s particularly useful when you need reliable swaps from solid input photos without building a large, complex graph.

よくある質問