Copy, Change, Publish: The Fastest Short Video for a Complete Beginner

Everything else in this playbook assumes you want to build something from a blank page. Most people starting out do not, and should not.

The fastest route to a short video that is genuinely yours is to start from one that already works, put your own character in it, and change it. No prompt writing, no tutorials, no understanding of how any of it works. A few clicks, and the result looks like a production team made it — because the settings behind it came from someone who knew what they were doing.

TJP’s own framing of this path is three steps: pick a look, add your talent, generate. This guide covers what actually happens at each one, and the two mistakes that turn “your own video” into “somebody else’s video you re-ran”.

Why this beats writing a prompt

A prompt is a description, and descriptions are the hard part. A beginner writing one gets a generic result, concludes the tool is overrated, and stops — usually within the first hour.

The recipe behind a good post is not a description. It is a prompt, a model and a seed, together, already proven to produce that result. Starting from one means the hardest decisions are already made, and you are choosing rather than guessing.

This is not a beginner’s crutch you graduate from. It is how experienced operators work too: nobody starts a new format from nothing when a working one exists to bend.

Step 1 — Find a look that is already working

Browse the community feed. It is readable without an account, and every post carries its counts: views, likes and recreates.

Sort your attention by recreates, not by likes. Likes measure whether people enjoyed it; recreates measure whether people wanted to build on it — which is the only signal that tells you a look is reusable rather than a one-off.

What to look for:

A style you could run repeatedly, not a single striking image. You are choosing a format, not a picture.
Something adjacent to an audience you can name. “Looks cool” is not a plan; “this is the look for the story I want to tell” is.
A post that allows recreating. Creators choose whether their work may be built on at all, and whether it costs credits. If it is closed, move on.

Step 2 — Recreate it, and understand what you just got

Press Recreate. The post’s recipe loads into your generator — the prompt, the model, the seed.

Three things are worth knowing about that moment:

The recipe was not public until now. Browsing shows the media, title and counts, never the prompt or seed — not in the feed, not in the page source. It is revealed to you when you recreate, and the attribution is recorded first.
The original creator is credited automatically, and may earn from it. That is the system working as designed, not something you are getting away with. See where the money comes from.
Everything stays editable. Nothing is locked. The recipe is a starting position, not a template you are stuck inside.

If the creator set a price, recreating spends credits and records the purchase — and recreating that same post again later does not charge you twice.

On your own work the button says Regenerate instead. Same mechanic; the difference is ownership.

Step 3 — The two changes that make it yours

This is the part people skip, and skipping it is the difference between having a video and having your video.

Change the seed. Always.

Recreate pins the seed, which means the default behaviour is the same picture again. That is deliberate — it is what you want when reproducing a result exactly.

It is not what you want here. Leave it and you have re-run somebody else’s video.

Clear or change the seed before you generate. That single change turns a reproduction into a fresh roll with the same proven recipe — the look you liked, a result that is yours.

(Some models manage their own seeds and do not accept one. For those there is nothing to pin and you get a new result either way.)

Add your talent

Swap in your own avatar. Now the character in the shot is your character, appearing in a look that was already working.

This is the step that compounds. A recreate with a stranger’s face is a one-off; a recreate with your locked character is an episode of something — because the next one has the same person in it, and the one after that. See building a channel persona for why recognition is the actual asset.

If you do not have an avatar yet, claim or train one before you build a habit around a face you cannot reuse.

Step 4 — Change one more thing, then generate

Now make it about your idea. Change the wardrobe, the setting, the time of day — one or two things, not everything.

Changing one element at a time teaches you what each part of the recipe was doing, which is how you stop needing recipes. Changing everything at once gives you a mess and no lesson.

Then generate. The cost is shown on the button before you commit.

Step 5 — Animate it, cut it, publish it

You now have a still you like. The rest of the pipeline is unchanged from the quickstart:

Animate the approved still into a clip, describing the motion and the camera. See prompting for video.
Repeat for each shot in your sequence — clips are seconds long, so an episode is many of them.
Cut them together, add sound and titles. See from clips to a finished episode.
Publish where the audience is, and pull a cover frame from the clip itself. See covers and thumbnails.

Copying a format versus copying a video

This distinction decides whether this approach builds you something or gets your account removed, so it is worth being exact.

Fine, and normal creative practice: adopting a format that is working — the pacing, the shot structure, the visual treatment, the type of story. Every genre in the history of film works this way.

Fine, and explicitly supported: recreating a post whose creator has allowed recreating, with attribution recorded automatically. The platform built this on purpose and creators opt into it.

Not fine: republishing someone else’s finished video as your own, or recreating with the seed unchanged and posting the identical result as original work. The first is a rights problem; the second is a duplicate, and platforms are good at detecting those.

There is also a practical reason beyond the ethical one. A copy cannot build an audience. Recognition — the thing that makes a channel worth anything — requires that your videos be identifiably yours. An exact reproduction of a trending post has nothing in it for a viewer to recognise next week.

And on studying what is trending elsewhere: watch the top short-drama accounts for structure, not for content to lift. How long is the hook? What happens in the first three seconds? Where does the episode cut off? Those answers are free, they transfer to any story, and using them is not copying anything. See writing a short-drama script.

The honest version of “make money fast”

Producing is genuinely fast. Copy, change the seed, add your character, generate — a publishable clip in well under an hour, on your first day, with no skill.

Earning runs on a different clock, and this path does not shorten it. An audience accumulates over weeks; platform payout thresholds take months; client work is the fastest route and still needs samples and a first conversation. See where the money actually comes from for realistic timelines.

What this path really removes is the reason most people quit in week one — the blank page, the bad first result, and the conclusion that you are not creative enough for this. That barrier is worth removing on its own. It is just not the same thing as being paid.