Video editing · Genjutsu 1.0
Genjutsu AI motion transfer & video editing
A performance is your starting point. Bring its movement into a new scene, or replace a subject in the footage you already have.
Your draft stays editable. Opening the model or a template does not start a generation.
Choose what you want to carry forward
Motion transfer
Use the source video as a guide to movement and timing, with reference images defining the subject or scene. This mode is suited to a performance, a character action or a product movement you want to reinterpret.
Object replacement
Start with a shot whose composition already works. Identify the person, prop or product to replace, and use the reference images to guide its new appearance. Review the surrounding scene as well as the edited subject.
For the dedicated fast object replacement mode at 720p or 1080p, see Genjutsu Lite 1.0. Compare the displayed quote for your actual source video before choosing.
From reference to reviewed video
Prepare the source
Upload a clip you have permission to use, or open an Explore template. Templates load their source video, ordered reference images and prompt together. You can replace individual images before generating.
Describe the edit
Use @Video for the source and @Image1, @Image2 and subsequent image references to identify your inputs. Keep the subject visible and the instruction specific. Prompts can be empty, and support up to 5,000 characters.
Review the quote and result
Choose a resolution and check the credit amount after the video finishes preparing. Sign in to generate. Completed results appear in Create and My Assets, where you can review, recreate and download them.
Genjutsu 1.0 settings and credits
The quote uses the measured source duration and selected resolution. The whole order is rounded up to a credit. Changing the source or settings changes the quote; simply preparing a draft does not spend credits.
| Resolution | Credits |
|---|---|
| 480p | 84 |
| 720p | 195 |
| 1080p | 558 |
Motion transfer rounds the measured duration to the nearest second, with a 4-second billing minimum. Object replacement rounds up each started second and accepts 4–30-second sources. The measured duration includes the longest audio or video track. Each generation returns one video.
Use your available Everygen credits. This model does not add a separate free allowance. See Everygen pricing for credit options.
Before you start
Do I need a separate photo for every angle?
No. One clear reference image can be enough to start. Additional images should describe the same intended subject or give useful visual context. Template images remain in their numbered order until you change them.
Can I use only a text prompt?
No. These modes need one source video and at least one reference image. For text-led video generation, choose another video model in Create.
What happens when I switch models?
Compatible inputs and your prompt stay in the draft. Everygen checks the new model’s settings and updates the quote. If references are incompatible, review the change before removing them.
Is Everygen the official Genjutsu website?
No. Everygen is an independent creative workspace. The controls described here are the Genjutsu 1.0 and Lite 1.0 options available through its integration; other services may offer different limits.