How to Use Genjutsu AI, Step by Step

The short answer: open the create page, pick a scene (or upload your own clip), add one photo, choose 16:9 or 9:16 and click Generate. A few minutes later you can download an MP4 where you replace the original performer — same moves, same camera, same background. This guide is about the choices that decide whether that result actually looks like you.

Step 1: Pick a scene

You can start from a ready-made effect or from your own footage. If this is your first video, start with an effect: each one has been tested with the built-in instruction, so the only variable is your photo.

Which effect to start with:

• Easiest: single-person scenes with a steady camera and the face visible most of the time — Street Runway and Red Suit Groove are good first picks.
• Harder: scenes where the performer turns away, spins fast or is small in the frame — the swap has less of your face to work with.
• Trend scenes: the Crazy Frog chase, Ozil goal and Flying Car pages each explain their own scene in detail.

The full effect library is described on Genjutsu Video AI.

Step 2: Choose your own clip (optional)

Uploading your own video gives you any scene you like, but the clip decides the result more than anything else. Before you upload, check that it has:

• One clear main person, larger than everyone else in the frame.
• One continuous shot — no hard cuts or transitions in the middle.
• The face visible for most of the clip, not just a back view.
• A steady or smoothly moving camera; shaky phone footage makes the edit flicker.
• Movement from start to end — 5 to 10 seconds of action works better than a long static intro.

Clips must be 3 to 10 seconds long.

Step 3: Pick the right photo

The photo is the only thing the AI knows about you, so it is worth thirty seconds to get right. A good photo is:

• Front-facing, eyes to the camera, head not tilted.
• Sharp at full size — zoom in and check the eyes are in focus.
• Evenly lit — daylight from a window beats a ceiling light or a flash.
• Unfiltered — beauty filters and heavy compression change your face shape.
• Uncovered — no sunglasses, masks, hands or hair across the face.
• A close match to the scene — if the performer has long hair or a hat, a photo with a similar look blends in more naturally.

Only use photos of yourself or of people who have agreed to appear in your video.

Step 4: Choose the format and generate

Pick 9:16 for TikTok, Reels and Shorts or 16:9 for YouTube and widescreen, then click Generate. If you are not signed in yet you will be asked to, and your scene and photo are kept so you can continue right after. The cost is shown next to the button before you submit; pack prices are on Is Genjutsu AI free?.

Step 5: Review before you post

When the render is finished it is saved to your videos. Before posting, play it once at full screen and check:

• the face in the fastest moments, where drift is most likely;
• hands and small objects, which can change shape between frames;
• the edges of hair and clothing against the background.

Common mistakes (and how to avoid them)

FAQ

Can I use it on my phone? Yes. It runs in the browser on phones and computers; there is nothing to install.

Can I make several versions with the same photo? Yes. Each version is a separate render, so you can try the same photo in different scenes, or a different photo in the same scene, and keep the one you like best.

Can I change 16:9 to 9:16 after rendering? No. The format is set before you generate, so pick the platform you are posting to first; a different format means a new render.

Ready? Pick a scene and upload your photo.