Teachers, trainers and presenters often explain things with their hands. Pointing out a direction, demonstrating a stretch, miming a lab procedure, showing how to hold a tool. Those movements carry a lot of meaning, but recorded footage of a person in a classroom or office is rarely engaging to watch twice. Motion transfer offers a way to keep the gesture and present it through an illustrated character instead.

This article explains how educators can do that with Genjutsu AI, a video-to-video tool designed only for motion transfer.

Why gestures matter in learning materials

When learners watch a demonstration, they follow the motion more than the person. The angle of an arm, the speed of a movement and the order of steps are what they need to copy. An illustrated character performing the exact same motion can keep all of that while matching the visual style of your slides or course, and it avoids putting a real person's face into every lesson.

How the process works

Genjutsu AI uses two inputs:

• A source video of 4 to 30 seconds in MP4 or MOV. This provides the motion and the camera movement.
• Up to 20 reference images. These provide the appearance, for example a course mascot or a simple illustrated presenter.

The output follows the timing and path of your recorded movement, with the look taken from the references. It does not generate video from text, so the demonstration always comes from a real recording.

Recording a demonstration that transfers well

• Show one step per clip. Break a procedure into short pieces rather than one long take.
• Face the camera at a helpful angle. If learners need to see a hand position from the side, film from the side.
• Keep the full movement in frame, including hands and feet if they matter.
• Use a tripod. For instructional content, a locked-off camera is usually clearest.
• Light the presenter evenly, avoiding a bright window behind them.

Designing the illustrated presenter

Create or commission a simple character and prepare reference images from the angles your clips use. If a clip shows a side view, include a side view in the references. Keep colours and clothing the same in every image so the character stays consistent across a whole module.

Budgeting for a course

Output costs depend on resolution:

• 480p: 3 credits per second, ideal for drafts
• 720p: 7 credits per second, fine for most course platforms
• 1080p: 11 credits per second, for projected or full-screen use

A module of eight 10-second demonstration clips at 720p costs 560 credits. Credits are bought in one-time packs, with no subscription, and they never expire, which suits course work that happens in irregular bursts. Failed generations are refunded automatically.

A note on privacy

Videos generated before your first purchase are shown publicly in the Community section. If your recordings include students or colleagues, get their consent and consider purchasing before rendering anything personal.

Scaling up

Institutions with development support can use the REST API to process a batch of demonstration clips with the same reference set, keeping a course's visual identity consistent.

Summary

Your gestures are the lesson. Motion transfer lets you keep them precisely while presenting them in a style that fits your materials. You can explore it at genjutsu.app.

— Genjutsu AI