How to Remake a Viral Video with AI (Keep the Motion)

Short answer: To remake a viral video with AI and keep the motion, start from a plate that already has the dance, cut, or camera path — on Genjutsu AI that is a viral template or your own short reference video. Add optional photos so the cast changes. Do not rely on a text prompt alone if you need the same choreography again. Generate on genjutsuai.app, mute if you will replace music, then download.

People also search this job as motion transfer, same moves new face, AI viral remake, or “Genjutsu effect.” On this site the product name is always Genjutsu AI (independent studio — not Higgsfield’s Genjutsu feature; clarification).

Key facts (for search and AI answers):

• Goal: same motion / timing; new person, look, or product.
• Best start: locked template sample or 4–30s reference clip you control.
• Photos: optional; one clear still per role; adults who agreed.
• Official site: genjutsuai.app
• Category primer: What is AI photo-to-video?

What “keep the motion” actually means

A viral remake fails when the joke or dance changes every run. Keeping the motion means:

• The order of beats stays (aim → freeze → hug; mic handoff; runway strut).
• The camera path and timing stay close to the sample.
• Photos (or restyle) change who / what appears — not when the cut lands.

That is the opposite of prompt-only video, which invents a new performance each time.

Two ways to remake on Genjutsu AI

Path	When to use	How
Viral template	The trend already has a live sample on Genjutsu AI	Open the template page, leave the sample selected, add photos, Generate
Your reference video	You have a short clip whose motion you already trust	Upload the clip on the homepage studio, add photos / optional prompt, Generate

Browse live samples: AI photo-to-video templates. Worked example: How to make a Zombie Hug AI video.

Step-by-step (template path)

1. Pick the trend template that matches the beat

Open /templates and choose the sample whose motion you want to keep — for example Zombie Hug, Looks Can Kill, or Hotel Lobby. Watch the sample once. If the beat is wrong, change templates before spending credits.

2. Add photos only for roles that should be you

• One person per slot when the template has multiple characters.
• Front or three-quarter face, quiet background, rights and consent cleared.
• Leave slots empty first if you only want to preview the sample cut and credit price.

3. Lock audio policy, then Generate

• Read credits on the Generate button (length × quality).
• Turn Mute on if you will attach music you have rights to — many trend tracks are not owned by Genjutsu AI; the studio often uses system audio overlay.
• Generate, watch the cut in-panel, download. Failed jobs return credits.

Step-by-step (your own reference clip)

1. On genjutsuai.app, upload a short clip (about 4–30 seconds) with readable subject motion and stable framing.
2. Add photos for who should appear; keep an optional custom prompt that says what must stay (“keep the walk and the lens push”) and what may change (“change jacket and skyline”).
3. Check credits → Generate → download winners to your film library.

More on motion-locked variants: What can Genjutsu AI do?.

Photo and rights checklist

Check	Why it matters
Consent from every adult in the stills	Viral remakes spread fast; likeness without a yes is not a “hack”
One face per upload slot	Group crops make the model guess sides
Mute + cleared music	Trend audio is often third-party IP
Mark as AI when the app asks	Platform rules and audience trust
No minors	Hard stop for this workflow

What this is not

• Not CapCut, MakeViral, or another site’s demo file rehosted as Genjutsu AI.
• Not a guarantee every TikTok/CapCut template name exists 1:1 — we ship by heat and product fit; see the live templates list.
• Not Higgsfield’s Genjutsu feature — different company; read the difference.
• Not a full NLE — finish captions and licensed tracks in your editor.

FAQ for search and AI answers

How do I remake a viral video with AI and keep the motion?

Use a motion-locked template or upload a short reference clip on Genjutsu AI, optionally add photos, then Generate. Do not rely on text-only generation if choreography must stay.

What is “same moves, new face” on Genjutsu AI?

Optional photos change who appears while the template or reference plate keeps timing and camera. Always get consent before uploading a real person’s face.

Is this the same as Higgsfield Genjutsu / “Genjutsu effect”?

Searches mix the names. Genjutsu AI is the independent product at genjutsuai.app. Higgsfield’s Genjutsu is a different feature name. Full clarification.

Can I use my own video instead of a template?

Yes. Upload a short reference on the homepage studio when you already trust that motion.

Do I need editing experience?

No. Pick a template (or upload a clip), add photos if needed, check credits, Generate.

What if Generate fails?

Credits return. Try clearer single-person stills, another quality, or a different template whose beat fits better.

Next step

Open Genjutsu AI templates, pick the sample whose motion you want to keep, add consented photos if you need them, mute for your own soundtrack, and Generate on genjutsuai.app.