Portrait restaging
Move the same person between locations, lighting setups, and camera treatments for campaign concepts.
Reference-anchored AI editing
Identity Preserving Image Editor
Edit your photos. Keep the people. Change everything else with a plain-English instruction in a free browser workspace.
krea2edit.app / live editor
The live editor runs in an embedded Hugging Face Space. If the shared GPU is busy, open it full-screen. Open the editor full-screen
What is identity-preserving editing?
Identity-preserving image editing changes a scene, outfit, object, pose, or visual attribute while using the source person as an identity reference. Unlike a general text-to-image regeneration, the goal is not to invent a similar person; it is to retain recognizable facial and character cues across the edit. Results still need human review because no generative model guarantees an exact match.
Identity-preserving editing use cases
The workflow is most useful when visual continuity matters more than generating a completely new subject.
Virtual wardrobe
Explore clothing silhouettes and colorways without losing the model who anchors the original photo.
Character continuity
Keep a recognizable protagonist across storyboard frames, key art, and visual development.
Local corrections
Replace props, recolor garments, or clean backgrounds while leaving unmentioned regions as stable as possible.
How it compares to regular AI editors
Both tools generate pixels, but they optimize for different outcomes. Choose identity-preserving editing when continuity is part of the brief.
| Regular AI editor | Identity-preserving editor | |
|---|---|---|
| Primary goal | Create a plausible edited image | Keep the referenced person recognizable |
| Face behavior | May regenerate or drift | Anchored to source identity cues |
| Best input | Text prompt or source image | Clear source person plus a focused instruction |
| Best use | Broad restyles and exploration | Restaging, try-on, composites, and continuity |
FAQ
Identity-preserving editor FAQ
What to expect before you upload a photo.
Is this the same as face swap?
No. Face swap usually replaces one face region. Identity-preserving editing can restage the whole image while using the source person as a reference.
Can it keep clothes and change only the background?
Yes, that is a strong use case. State only the new background and avoid extra style instructions that could alter clothing.
Can I use more than one person?
Single-subject images are the safest starting point. Multi-person composites can blend identities, so build complex scenes in controlled passes.
Is the editor free and private?
The embedded community Space is free to try. Your file is processed by that third-party Space, so do not upload sensitive images and review its current terms.
Make the edit without recasting the person
Start with one clear portrait, describe one change, and compare the generated result with your source.