An AI fashion lookbook video is built from stills, not from footage: you photograph or generate each look as a single frame, then animate each frame separately with a motion effect — a runway walk, a catwalk turn, an outfit reveal — and cut the clips together, which on VdoBloom costs roughly 17 to 82 credits per look depending on the model and resolution you choose. Nobody books a studio, a runway or a crew. The constraint moves from production to art direction.

If any look is animated from a photo of a real person — a model, a friend, a client, yourself — you must have that person’s consent to use their likeness in generated video before you upload the photo.

Why stills-first beats text-to-video for lookbooks

A lookbook has one job: show the garment accurately. Text-to-video models invent clothing. Ask one for “a model in a sage linen jumpsuit” and you get a plausible sage linen jumpsuit, not your sage linen jumpsuit — wrong buttons, wrong drape, wrong hem. Image-to-video models animate what you give them, and preserve it far better.

So the pipeline is: one hero still per look, then motion. This also solves the consistency problem that wrecks most AI fashion content. The same face, the same styling and the same lighting carry across every clip because they all originate from the same still or the same photoshoot, rather than being re-rolled by the model each time.

Step 1: get one clean still per look

Two routes work. If the garments exist, photograph them on a person or a mannequin in flat daylight, full body, feet in frame, plain wall behind. Then fix the background and light in the image editor. If you are working from flat-lay product shots or want to preview styling before sampling, the AI fashion try-on tool puts a garment onto a model image instead.

Three things to check on every still before you animate it, because motion amplifies all of them:

Feet and hands fully in frame. Cropped limbs are where image-to-video models invent the worst geometry.
Clear separation from the background. A dark garment against a dark wall will smear during motion.
A neutral, upright starting pose. Extreme poses give the model less room to move plausibly.

Step 2: pick the motion for each look

A lookbook that uses the same movement fifteen times reads as a template. Vary it. VdoBloom ships several purpose-built photo-motion effects, each with real templates and tuned default prompts:

Effect	Templates available	Best used for
Fashion Walk	Classic Runway, Street Style, Evening Gown, Casual Chic	The opening and closing looks; full-garment movement
Catwalk Turn	Runway, Fierce, Elegant	Showing the back of a garment without a second photo
Outfit Reveal	Fashion, Dramatic, Elegant	Transitions between looks; the hero shot of a drop

The Fashion Walk templates are the most editorially distinct. Classic Runway gives hard overhead light, a glossy reflective floor and a locked-off frontal camera. Street Style moves to overcast daylight with a tracking camera and shallow depth of field. Evening Gown is a slow, low-angle rise with warm specular highlights. Casual Chic is a steady eye-level mid-shot with bright natural light — the closest to a genuine lookbook rather than a show.

Every one of those default prompts contains explicit preservation language instructing the model to keep facial features, body proportions, clothing, pose, lighting and identity exactly as in the source image. Do not delete that when you customise — add your direction after it.

Step 3: choose the model and know the cost

Credits are the same currency across the platform, and 1 credit is $0.05 at the Lite tier ($15 for 300 credits). Real per-generation costs for the common image-to-video choices:

Model	Tier	Credits	Cost	When to use it
Seedance 2 Mini	720p, 5s	17	$0.85	Drafting every look cheaply before committing
Seedance 2 Mini	720p, 10s	33	$1.65	Longer walk-and-turn in one take
Seedance 2 Fast	480p, 5s	24	$1.20	Motion test only, not for delivery
Seedance 2 Fast	720p, 5s	50	$2.50	The practical everyday lookbook clip
Seedance 2	720p, 5s	82	$4.10	Hero look, best motion quality
Seedance 2	1080p, 5s	203	$10.15	Campaign-grade single shot
WAN 2.7 image-to-video	720p, 5s	32	$1.60	Cost-efficient mid tier
WAN 2.7 image-to-video	1080p, 5s	48	$2.40	Cheapest honest 1080p delivery
Kling V2.5 Turbo	5s	40	$2.00	Clean, stable human motion
Kling V2.5 Turbo	10s	80	$4.00	Full walk with a turn at the end
Veo 3 Fast	—	40	$2.00	When you want native audio in the clip

The exact cost for your selected model, duration and resolution is shown before you press generate, so nothing here is a surprise charge. Note how steep the 1080p Seedance 2 jump is — 203 credits against 48 for WAN 2.7 at the same resolution and length. For a twelve-look lookbook that is the difference between $28.80 and $121.80.

A worked twelve-look budget

Draft all twelve looks on Seedance 2 Mini at 720p 5s: 12 × 17 = 204 credits. Expect to re-roll about a third of them once, so add four more: 68 credits. Then re-run the eight keepers plus four fixed drafts at WAN 2.7 1080p 5s: 12 × 48 = 576 credits. Total roughly 848 credits, about $42, for a full 60-second lookbook cut. Doing it entirely at the flagship 1080p tier would cost more than triple that for motion quality most viewers will not notice at 9:16 on a phone.

Step 4: cut it like a lookbook, not like a reel

Generated clips are short by nature, which suits the format. Hold each look for three to four seconds, cut on the movement rather than at the end of it, and keep the camera grammar consistent — do not mix a tracking street shot with a locked runway shot without a reason. Export vertical 9:16 for social and a 16:9 version for a site hero from the same stills if the aspect ratio is available on your chosen model.

Two practical warnings. First, fabric physics is where these models still give themselves away; long trailing hems and sheer layers move less convincingly than structured garments, so put your tailoring at the front of the edit. Second, never use a lookbook clip as the sole representation of a garment on a product page — the still is the truth, the video is the mood.

Rights, watermarks and the free tier

Free accounts get 10 credits with no card, which is enough for one draft clip — useful for judging whether a still animates well at all before you invest. Commercial use of a lookbook means a paid plan: paid downloads are watermark-free and carry commercial rights. Plans and the one-time credit packs that never expire, from $2.49 for 75 credits, are listed on the pricing page, and annual billing brings Lite to $10 a month.

Frequently asked questions

Do I need a real model or can the whole lookbook be generated?

Both work. If you use a real person’s photo you need their consent for AI-generated video of their likeness. If you generate the model image instead, use the same generated person across every look so the lookbook stays consistent, and disclose that the imagery is AI-generated where your market requires it.

How long can each clip be?

It depends on the model. Seedance tiers run from 4 seconds upward in one-second steps, Kling V2.5 Turbo offers 5 and 10 seconds, and WAN 2.7 image-to-video covers 5, 10 and 15 seconds. Three to four seconds per look is usually the right edit length regardless of what you generate.

Will the garment change during the video?

Mostly no, if you keep the preservation wording in the default prompt and start from a sharp full-body still. Drift shows up most in small details — buttons, prints, jewellery — and in long flowing fabric. Shorter clips drift less, which is another reason to generate 5 seconds rather than 10.

Which effect should I use for showing the back of a garment?

Catwalk Turn, using the Runway or Elegant template. It pivots the subject rather than walking them away from camera, so the reverse of the garment reads clearly in a single short clip without needing a second photo shot from behind.

What is the cheapest way to test whether a still animates well?

Run it once on Seedance 2 Mini at 480p for 5 seconds, which is 8 credits or $0.40. If the limbs and fabric hold together at 480p, they will hold at 720p and 1080p. If they fall apart, fix the still rather than paying for a better model.

How to Make an AI Fashion Lookbook Video on VdoBloom