Film noir
Jaw edge under hard shadow
Six real proof runs · all visible
Keep the shot. Replace only the identity.
Target pose · light · wardrobe+Source identity→Result same scene, new face
Technical light
Helmet reflections stay target-owned
Painted portrait
Canvas texture survives the swap
Campaign portrait
Camera height and expression align
Warm kitchen
Skin temperature matches the scene
Fantasy character
Hair and wardrobe remain intact
30-second preflight
Match the inputs before spending a credit
Our failed runs were usually predictable before generation. A clean source cannot rescue a target with an incompatible angle, expression, or lens.
- 01YawKeep front, three-quarter, or profile views close.
- 02PitchMatch camera height—especially phone selfies.
- 03ExpressionAlign mouth, squint, brow, and cheek tension.
- 04OcclusionCheck hair, glasses, hands, helmets, and shadows.
- 05LensSimilar framing avoids stretched or compressed features.
PhotoOne identity solveBest for campaign stills and portraits
VideoTrack every frameMust survive motion, cuts, blur, and occlusion
A good still swap is not a video workflow
The photo tool above returns one edited image. A moving target needs identity reacquisition and temporal consistency; independently swapping frames creates flicker. Review video at normal speed and frame by frame, or change the whole character when identity is only one part of the brief.
Delivery gate
Permission is part of the input
Your face, a contracted model, or an adult who approved this exact context.
Label synthetic identity when viewers could mistake it for a real event or endorsement.
No minors, sexualized edits, fraud, harassment, identity documents, or deception.
Retain source portraits only as long as the approved project requires.
If the subject would be surprised by the finished context, stop and get consent.
Why source and target order matters
The target owns pose, body, wardrobe, lighting, and background. The source supplies facial identity. This Voor tool is intentionally image-only and should not be confused with Higgsfield's separate video product.