How to remake a meme video with AI

What a remake actually is

The original trend clip is a reference for the format: where the camera sits, how many people are in frame, what happens on each beat. The model does not copy its pixels. It reads a written description of that format plus your photos, and generates a new 4-to-12-second clip.

That is why the output keeps your face, your hair and your clothes. It is also why the background can look slightly different from the original: the model is recreating the scene, not compositing it.

Step 1: pick a format, not a video

Start from a template page. Each page lists the photo count, the aspect ratio, the render length and a shot-by-shot description. If you cannot find the exact trend, open the studio and describe it yourself: who is in frame, what they do, where the camera is, and what changes on the last beat.

A good prompt is concrete. “Two friends hold hands and do a two-step on a sunny street, then laugh at the camera” gives the model a location, two people, an action and an ending. “Make a funny dance video” gives it nothing to hold on to.

Step 2: choose photos the model can read

One clear, front-facing photo per person beats a group photo every time. The face should take up at least a third of the frame, with even light and no heavy shadow across the eyes. Sunglasses, hats and glasses are fine — the model keeps them.

For dance formats, use a knees-up or full-body photo so the model can see the outfit and shoes. For reaction formats, a head-and-shoulders photo is enough. Do not upload photos of minors or public figures: the content check refuses them before any render starts.

Step 3: match the render size to where you will post

9:16 fits TikTok, Reels and Shorts. 16:9 fits YouTube and X. 4:3 and 1:1 are for formats that were shot that way. The aspect ratio is part of the template; changing it changes the framing the prompt describes, so use the one on the page unless you are writing your own prompt.

A 4-second clip costs 10 credits, 8 seconds costs 20, and the default 12-second render costs 30. Duration scales the price linearly and rounds up to a whole credit.

Step 4: pay once or use a plan

A single 4-second video is $9.90; the 12-second render is $19.90 once. If you plan to make several, the monthly Starter plan gives 20 credits for $9.90, and Pro gives 200 for $39.90. Credits are charged only when a render succeeds; a failed render returns them automatically.

You do not need an account. Enter your email at checkout, pay, and the finished video is sent to that address with a private link. You can sign in later with the same email by one-time code to find it under My Videos.

Step 5: check the result before you post

Watch the whole clip once at full size. Look at the face in the middle and at the end, not only the first frame — drift usually shows up in the last two seconds. Check the hands and the background edges, where generation artefacts are easiest to see.

If the face drifted, re-run with a sharper photo of that person. If the motion is wrong, edit the prompt: name the action, the camera move and the ending. Re-running costs another render; the first one is not refunded unless it failed technically.

Why a render sometimes fails

The most common causes are a photo the model cannot read, a prompt that describes something the model will not generate, or a provider timeout. A technical failure refunds the credits. A result you simply do not like is not a failure — treat it as a prompt problem and change one variable at a time.

What the remake cannot do

It does not add the original song. Renders are silent; add the trending sound in the app you post from. It does not guarantee a perfect likeness from a blurry photo. And it does not remove the need for consent: only upload people who agreed to appear, and label the video as AI-generated where the platform asks.

Try it on this page

FAQ