Edit a Person in Existing Video

It modifies an existing video using a reference image and edit instructions, producing an edited video. The recipe is intended for changes to subjects, elements, or visual styles while preserving the original motion and composition.

It can help social and content teams, filmmakers, entertainers, and artists explore focused changes to footage they already have. It is most relevant when the desired result is an edited version of an existing video rather than a new video from scratch.

Prepare an original video, a reference image, and a clear edit instruction. The saved example uses “the person” as its instruction; in Fuser, you can change the exposed video, image, and text inputs for a different edit.

The saved example shows the original video, reference image, and “the person” instruction connected to an edited-video output. It represents the recipe’s saved default example, not a live generation on this page.

Yes. The exposed edit-instruction input can be changed in Fuser, and you can also provide a different reference image or original video. Instructions can describe a subject, element, or visual style you want to modify.

The recipe produces one edited video. Its purpose is to demonstrate a requested change while retaining the original footage’s motion and composition, though the specific result depends on the inputs and instruction used.

It suits social content, film and entertainment work, and art or illustration explorations involving existing video. It is useful for testing or presenting alternate visual directions, rather than for unrelated still-image tasks.