You have a target image. You have a source face. You want the source face on the target cleanly, quickly, with no manual masking or complex setup.
Two images in. Face-swapped result out.
Run it now on Floyo!
How It Works
Upload your source face image and your target image. ReActor detects the face in the source, maps it onto the face in the target, and runs a CodeFormer restoration pass to clean up the result. You get a before/after comparison and a saved output automatically.
No prompts. No settings to tune for most runs. Two images in, one clean swap out.
Key Inputs
Source Face Image
The face you want to transfer. A clear, front-facing photo with good lighting gives ReActor the most data to work from.
Works well with:
clean portrait photos
front-facing or slight angle shots
well-lit images with no heavy shadows over the face
high-resolution sources with sharp facial detail
Works less well with:
heavily obscured faces (sunglasses, masks, extreme angles)
very low-resolution or blurry source images
group photos where multiple faces are present
Target Image
The image you want to apply the face to. The source face maps onto the detected face in this image.
Works well with:
any image with a clearly visible face
portraits, character art, generated images, product shots with a model
single-face images for the cleanest result
For multi-face target images, ReActor detects all faces. It will apply the swap to the first detected face by default.
CodeFormer Restoration
Runs automatically after the swap. It restores facial structure and sharpens detail that can get soft during the face replacement process. The restoration is what separates a clean swap from an obvious one.
What This Is Great For
Character consistency: Apply a specific face to AI-generated characters across multiple images. Keep the same identity across a series without regenerating from scratch.
Creative and concept work: Swap faces for storyboards, pitch decks, concept visualizations, and creative projects where you need a specific likeness in a specific scene.
AI image finishing: Generated images often produce generic faces. ReActor lets you replace them with a specific reference face for more controlled output.
Rapid iteration: The workflow runs in about 95 seconds. Fast enough to test multiple source faces against the same target image in one session.
What to Watch Out For
Only swap faces you have the right to use. ReActor is a powerful tool, use it responsibly. Don't use it to impersonate people without consent or to create misleading content.
Source face quality is the most important factor. A blurry, low-resolution, or heavily shadowed source face produces a weak swap. Use the clearest, sharpest portrait you have.
Extreme angle mismatches between source and target produce less accurate results. A front-facing source face swapped onto a heavily turned head will look off. The closer the head angles match, the cleaner the result.
CodeFormer restoration runs at a fixed strength in this simple workflow. If you need more control over restoration intensity useful when the restoration is smoothing out too much detail a more advanced ReActor workflow with tunable CodeFormer settings would serve better.