StyleRoom Insight · Guide

Style Clip

How to write a great AI fashion video storyboard!
Write it like this
to get exactly the scene you want.

Slow motion, hair in the wind, how many outfits, any move: what to film is entirely up to you.
This guide shows how to spell that idea out so the video AI understands it.

Published
Reading time 5 min
By StyleRoom Video Team

In Style Clip, you can make a video from just your own photos and an idea, without picking a template, right?
But when you wrote only one line, you may have thought, "Hmm, that's a bit different from what I pictured."
The StyleRoom video team gathered the storyboard tips we learned while building video templates and checking the real results.
We'll show real Style Clip template videos along the way!

At a glance
  • PrincipleVideo AI draws what's written, and fills in what isn't on its own.
  • TimeOne move, glance or reaction every 0.5–1 second, with start time and how long it holds.
  • SceneWrite moves, outfit, camera, place and background motion in words you can see.
  • HelperEven a short note gets expanded to this level of detail by the storyboard helper.

01 · BasicsGood to know first

Video AI draws what's written exactly, and fills in what isn't on its own.
Those filled-in parts are where a video feels "flat" or "not what I imagined."

What to setWritten brieflyWritten in detail
TimePose for 10 secondsLooks back at 1.5s and holds for 1 second
MoveA cute poseRight-hand V sign beside the cheek, head tilted slightly left
Camera(not written)A friend holds it at waist height, full body in frame
PlaceAn alleyRed brick pavement underfoot, a white low-rise and a utility pole behind
Light(not written)4 p.m., warm sunlight coming in low from the left

02 · TimeTime: one thing every 0.5–1 second

This makes the biggest difference.
If you give a long stretch just one move, the AI stretches or rushes the timing.

Templates with a first-second hook, quick beats and a planned ending

03 · MoveMoves: which part, how much

In each move line, include which body part + which direction + how much + the ending pose.

BriefStrikes a lively pose.
DetailedBends one knee back to kick the foot up lightly, keeps her balance with a playful face, and puts it down about half a second later.

Templates where moves with a set hand, direction and size follow one after another

04 · OutfitOutfit: move so it shows

If the video is about the clothes, look at the outfit in your photo and pick moves that show that outfit well.

What to look for in the photoExample moves
MaterialLight fabric settles slowly after a turn, and a heavy coat's hem swings weightily with each step
Length and fitA long coat's hem follows half a beat behind when walking, and a cropped top shows the waistline when the arms go up
DetailsPleats fan out and close on a turn, hands go into pockets if there are any, and a collar gets straightened by hand
The side that catches lightBy a window, turn toward the light so the fabric texture shines

Templates where the outfit moves in a way that suits its material, length and details

05 · CameraCamera: who films, and how

Without camera notes, the AI films like a slick ad or zooms anywhere.
Set it in one line at the start of each scene.

The camera circling the model, and a fixed camera with the model turning

How the frame changes with who films where: a mirror selfie and a street shot

06 · PlacePlace, light and small background motion

One line for the background gives you a cramped, awkward space.
Go in the order underfoot → just behind → far background → top of the frame, with each object's material, color and position.

BriefA street corner in Seoul.
DetailedRed brick pavement underfoot, double yellow lines along the curb, a silver car parked across the street, a tall tree with light green leaves, a white low-rise with dark windows, power lines passing overhead.

A frozen background looks like a person pasted onto a photo.
Pick 2–3 things already in the scene and let them move a little.

PlaceWhat can move
Streets and alleysCars and bikes passing blurry in the distance, people walking, swaying trees and power lines
Parks and terracesLeaves and their shadows, passing cloud shadows, birds flying by
Waterside, stations, airportsGlittering ripples, trains and travelers passing in the distance
IndoorsDust in the window light, gently shifting light, curtains and plant leaves already there

Writing size and distance together, like "small and blurry far behind the model," keeps them from covering the subject.
Setting a time, like "around 6s a cloud passes and the shade comes and goes," also gives the video a sense of time passing.

Passing cars and people, and light from one side, bring the scene to life

07 · OutfitsWhen you have several outfits

Three ways to switch outfits by pairing image numbers with segments

08 · WordsPhrasing that works well

Video AI draws words as they are.
Even if you write "does not ...," the word itself can make the thing appear.

If you writeTry this instead
No other people appearThe model is alone
The camera does not shakeThe camera is fixed
Without cutsOne continuous shot from start to finish
Pretty, cool, emotionalVisible moves, light and place
Swings her armCovers the lens with her palm (make the purpose visible)

The only "does not" you need is the final line: "No text, subtitles, logos or watermarks."
The aspect ratio is chosen in the app, so you don't need to write it.

09 · ExampleOne line → storyboard example

We expanded the one-liner "walking in front of a café with a coffee, then looking back" using these tips.
The idea stays the same; only the detail goes up.

Once at the start: camera · place · light · background motion Camera: A friend films with a phone held at waist height. Front view, full body from head to toe, natural slight handheld shake. Place: Gray paving blocks underfoot, a glass café window just behind with a small wooden bench in front of it (left), a dark green awning at the top edge. Light: Warm late-afternoon sun coming in low from the right, long shadows stretching across the pavement. Background: A bicycle passes blurry far behind the model, and the edge of the awning sways very slightly. Scenes 0–3s: Holding a takeaway cup in her right hand, she walks slowly toward the lens. With each step the coat hem follows a few centimeters behind. Around 1.5s the wind blows her bangs and she tucks them behind her ear with her left hand; at 2.5s she looks straight at the lens and smiles. 3–6s: Around 3.2s she takes a sip and stops in front of the bench. Around 4.5s she follows a passing bicycle with her eyes and smiles to herself. Around 5.5s she slips her left hand into her coat pocket. 6–9s: Turning her back to the camera, she takes two steps toward the café door. The coat swings heavily with her steps and settles back into its shape in the photo. At 8s she glances back sharply over her right shoulder. 9–10s: Still looking back, she meets the lens and lifts the cup slightly, and the video ends on that pose. Final line Keep the face, body and outfit as in the photo. No text, subtitles, logos or watermarks.

Compared with the one-liner, more than 10 moves are in, each with a time.
Moves tied to the outfit like the coat hem, a small reaction to the bicycle, the moments she looks at the lens and the ending are all set.

TIP. Writing at this level every time takes effort, right? Even a short note gets expanded like this by the storyboard helper. Just write down the scenes you want to keep!

10 · CheckBefore you generate

11 · FAQStyle Clip storyboards: frequently asked questions

Do I have to write a long storyboard?

No.
Even a short note gets expanded by the storyboard helper.
But if there's a scene you must keep, write that part yourself.
It will be kept as written.

Should I describe the face or clothing colors again in text?

No.
The face, body, clothing colors and logos are already in your photos.
Describing them again can clash with the photo if it's even slightly off, so just name the material and shape briefly.

Can I ask for slow motion or hair blowing in the wind?

Yes.
Just write the effect you want in that segment.
Put the segment and the size together, like "3–7s: this whole segment is slow motion, the hair blows strongly in the wind."

When should I write "does not ..."?

We recommend using it only in the final line: "No text, subtitles, logos or watermarks." Everywhere else, describing what you want, like "the model is alone" or "the camera is fixed," works better.

It draws what you write. Write what you want to see.

Make your first video from your own storyboard in Style Clip today.

Style Clip · Storyboard
Make a video from my storyboard
Even a short note gets expanded by the storyboard helper.
Try it now →