Clone any video into yours

Paste a link to a video that is already working in your market. BrandU takes it apart into eight structure parts: script, shots, captions, voiceover, B-roll, effects, pacing and format. It rebuilds each part with your product, your claims and your voice. You edit the script instead of a timeline, and you choose the aspect ratio.

By Elias Sun, Founder of BrandU

What cloning actually copies

Cloning copies the structure of a video that sells, not its pixels. The structure is the part that took the money and the testing to find: the order of the argument, where the hook lands, how long each beat holds, how the captions come in. The footage, the product and the words are the parts you replace.

BrandU reads the reference the way an editor would. The pipeline takes a screenshot every second of the video, studies the whole sequence, and writes down what it sees as a production sheet: what each shot does, how long it runs, what the captions say and when they appear. That sheet is the structure. It is data, not a copy of the picture.

Starting from a blank prompt is a different kind of bet. You are asking a model to invent a structure, a pace and a hook at the same time it renders your product, and most attempts lose on one of the three. Starting from a video that already works removes that guess. The structure is proven before you begin; your job is to say what is yours.

A video broken down into a structure sheetThe pipeline takes a screenshot every second of the reference video: 30 seconds becomes 30 frames. Each one is recorded as what the shot does, how long it runs and when the captions appear. That record is the eight-part structure you rebuild with your own footage, product and words.A VIDEO THAT WORKSa screenshot every secondWHAT BRANDU READS00:0000:0300:1200:21a production sheet — the structure, as dataYou keep this: order · hook · pacing · captionsYou replace this: footage · product · words
Structure is what carries over. Footage, product and words are what you replace.

The eight parts, and what replaces each

A video works because of how its parts fit together, so that is what gets taken apart. BrandU breaks the reference into eight structure parts, and you rebuild each one with your own material.

Here is what that looks like on a real example. Take a 30-second UGC ad for a skincare product. The order of the argument carries over: problem, then the product in use, then the objection, then the offer. Where the hook lands carries over, three seconds in, before the first product shot. How the captions come in carries over, and so does the cut rhythm. What does not carry over is anything that identifies the original: the person, the jar, the claims, the brand name.

That is why the pages that clone well are the ones with a format already working for them. UGC talk-throughs, product reviews and multi-host interviews are all structures with a repeatable shape. Ecommerce teams use it to put a whole catalogue through one format. Local businesses use it for one format per location. Creators use it for a channel that publishes on a schedule.

There is a second half to that answer, and it is where most people get it wrong. What carries over is not only the order of the argument. It is also how the shots are composed, where the light comes from, and the overall look of the frame — the method behind the video. What does not carry over is the work itself: the footage, the people, the words and the brand. That line is deliberate. A style can be borrowed and still belong to you; a scene cannot.

That holds whether the reference is thirty seconds of product footage or the opening of a film trailer. A brand team can take the pacing and the light from a trailer it admires and rebuild the beat around a launch. A short-drama team can take the hook rhythm of a teaser that already holds viewers and swap in a new episode. The starting point is always the same thing — a video that already works.

DimensionKept from the referenceRebuilt with your content
Script structureThe order of the argumentYour product and selling points
ShotsHow shots connectYour footage and product
CaptionsWhen and where they appearYour copy
VoiceoverTone and rhythmYour voice
B-rollWhere it cuts inYour clips
EffectsWhen they fireYour brand's look
PacingHow much happens per secondWhat you emphasise
FormatAspect ratioYour platform

How closely to clone: keep the shot plan, the structure, or just the idea

The same reference can be rebuilt at three different distances, and the choice is yours to state in one line. The only difference between them is how many of the eight structure parts you keep. Keep more and the result sits closer to the original. Keep fewer and more of it is yours.

Keep the shot plan. The shots arrive in the same order, the cuts land in the same places, and the rhythm is the one that already works. What changes is the product, the script and the brand. Pick this when the reference is close to the thing you want to make, and the gap between it and your product is mostly the content inside the frame.

Keep the structure. The argument keeps its order and its pace — the problem, the proof, the objection, the offer land where they landed. The shots themselves are rebuilt, and the look is yours. Pick this when the reference proves a way of arguing that you want, and your footage would never have been shot the way theirs was.

Keep only the idea. One thing carries over: where the hook lands, and how fast the first few seconds move. Everything after it is rewritten for your product. Pick this when the reference is a signal that a format is landing in your market, and you want to answer it in your own voice rather than in its shape.

None of the three is more advanced than the others, and none of them costs a different amount. They are the same machine pointed at different targets — which is why the choice lives in a sentence of your brief rather than in a settings panel.

The whole video is generated at once

A video here is made as one piece. The shots, the captions, the voiceover and the pacing are parts of the same job, not separate pieces to be finished one at a time and joined afterwards.

That difference shows up before the render, not after it. When each shot is generated on its own, a longer video becomes a chain of generations, and the joins are where the look drifts, the voice changes and the captions stop matching. When the video is generated as one piece, there are no joins to fix.

It also means length is not a setting. What you write is what renders, and a short ad and a longer video are the same job at different sizes.

One video becomes a workflow you keep

The result of a clone is not a finished clip. It is a workflow: a stored structure you can keep editing and keep reusing, with your product and your voice already in it.

The difference shows up on the second video, not the first. A finished clip is used once and then it is spent. A workflow is the thing you come back to next week: you open it, change what needs to change, and render again. The argument has already been settled, so next week's work is not "what do we say" — it is "which part of this proven thing do we vary now".

That is where the volume comes from. One reference video gives you 100+ variants, because every hook, product, opening or language you swap in is a variant of the same structure. Enough to test the angles instead of guessing which one wins.

The other half of that is where the work happens. Reading the reference, writing the script, changing what you do not like and rendering the finished video all sit in the same workspace. You do not export a breakdown to one tool, rebuild it in another and hand the result to a third. There is nothing to hand over, because the structure the reference was read into is the same object you edit.

Your brand, your presenter, your products and your footage also stay in that account, so the next video starts with them already in it. The tenth video is not the first video done again — it is the first video, updated.

Change it a hundred times, pay for the change

Most video tools charge you the same way every time you touch a file: the model runs again from the start. BrandU is built so that most edits do not need the model at all.

What you pay for depends on what you changed. Reorder the shots, restyle an element or swap an asset, and the model does not run again — only the render repeats. Rewrite one segment of the script, and only that segment and its voiceover run again, plus the render. Rewrite the whole script, and everything runs again. Nothing is charged before you generate.

So the tenth revision of a workflow costs less than the first, and the hundredth costs less than the tenth. Competitors cannot follow you here, because every edit they make starts the model over.

This is also why editing feels different. When you change a line, BrandU re-times the elements around it, so the captions, the voiceover and the cut points follow on their own. You are not dragging clips to keep them in sync; the structure does that for you.

Start from a link, or from a sentence

You can start from a video you have seen work, or from nothing but an idea. Both entry points end in the same place: a workflow you can edit.

  1. 1

    Paste a link

    Paste a video link. A page URL is fine, and no file is needed. With no link yet, describe the video in one sentence.

  2. 2

    Make it yours

    Replace the script and the on-screen product, and set the brand voice. The shots, the pacing, the captions and the voiceover follow your words.

  3. 3

    Render

    Generate the variants and download the finished videos from the tasks page.

If you have no video to clone yet, describe the one you want in a sentence. BrandU drafts the script from scratch, and everything after that is the same: you edit the script, you render, and you keep the result as a workflow. The two paths meet in the same editor.

Video cloning FAQ

Does cloning make an exact copy?
No, and it could not. BrandU keeps the structure — the running order, the pacing, where the hook lands, how the captions come in — and replaces everything that identifies the original: the footage, the product, the claims and the voice. That is deliberate. An exact copy of someone else's ad is their ad, and it would not say anything about your product. What you get is a video built on a shape that is already known to hold attention, in your own words.
Can I clone just the captions or the opening?
Yes. You can reference any part of the structure: the caption style, the opening few seconds, where the B-roll cuts in, or the pacing. This is useful when you already have a workflow you like and only one element is failing — a weak hook, or captions that arrive too late to be read. Change that part and leave the rest alone; only what you changed runs again.
Can I use my own product and presenter?
Yes. Upload your product and your assets, and BrandU rebuilds the script and the shots around them. Your own footage works as B-roll, product images work as the on-screen product, and the voiceover is set to your brand voice. If you have a presenter you use consistently, describe them once and every segment reads from that description, so they stay the same person across the whole video.
Can I make several versions from one video?
Yes — that is what the workflow is for. One reference video gives you 100+ variants, because every hook, product, opening or language you swap in is a variant of the same structure. Teams use this to test angles rather than guess: same proven shape, a different opening line, and the numbers tell you which one to keep. The workflow you started from is still there afterwards.
Do I need editing experience?
No. There is no timeline and nothing to drag. You work on the script: you rewrite lines, reorder the shots, and swap the footage. When you change a line, BrandU re-times the elements around it, so the captions, the voiceover and the cut points follow on their own. If you have never opened an editor, that is the intended workflow — you describe what you want in words, and the video follows the words.
How do I copy a video and make it mine?
Paste the link. BrandU reads the video, takes it apart into the eight structure parts, and gives you a script built on that structure. You rewrite the script with your product and your voice, adjust any shot you want to change, and render. The result is yours — it uses the shape of the video you started from, and none of its content.
How close to the original can the result be?
As close as you want to name it. You can keep the shot plan and rebuild the content inside the shots, keep the structure and rebuild the shots themselves, or keep only where the hook lands and rewrite the rest. What never carries over is the original's footage, its people, its claims and its brand — not because the tool cannot, but because a copy of someone else's ad is still their ad.
How long can a finished video be?
There is no fixed cap. The video is generated as one piece, so length follows your script rather than any per-generation limit, and nothing is joined afterwards. In practice the constraint is the story: how much video a script is worth writing.

Start from what already works

Clone a video that already works, rewrite the script with your product, render in minutes.