Blend a source face into image or video targets with a cleaner internal workflow, GPU-first defaults, and storage-backed results.
Mode selection
Image Swap uses a source image and target image. Video Swap keeps the same source image flow and swaps onto a target video.
Credits
--
1 credit per successful generation
Locked execution defaults
Face swapper weight
0.65
Execution
cuda / 12 threads
Memory
strict / 24 GB
Mask + encoder
blur + occlusion
Upload the face reference image that should drive the swap.
Upload a reference image to preview it here.
Upload the target image that should receive the swapped face.
Upload a reference image to preview it here.
Processors
These selections are applied through a per-job FaceFusion config so the working local CLI flow stays stable.
The current installed FaceFusion CLI reports processor choices inconsistently, so AnotherWrapper applies these through the generated job config instead of direct `--processors` flags.
Preview
A lightweight side-by-side preview helps verify your source asset and target media before generation.
Source
Target
Upload a reference image to preview it here.
Generate
The current workflow keeps auth, credits, and storage-backed outputs intact while mirroring the working local Windows setup.
Current run
Image Swap
face_swapper + face_enhancer
Result
No output yet
Upload your source and target files, choose processors, then generate a final image result.