Demo library · 8 fictional brands
One brief. A whole campaign.
Each brand below started as a single short brief. Everything after that was generated end to end: brand kit, key visuals, social posts, a UGC video with voice, a 15-second spot, a story ad and the landing page you can click into.
Fictional brands. All imagery, voices and video are AI-generated illustrations, not customer results.
- The brief
- 01Brand kit
- 02Key visuals
- 03Social posts
- 04UGC video with voice
- 0515 s spot
- 06Story ad
- 07Landing page
The library
Hover a card to preview its UGC ad (muted). Open one to see the full campaign on its own bespoke landing page.
ConceptThe Blue Hour Shift
“Baked while the city sleeps.”
ConceptThe Lime Line
“Follow the line. Land softer.”
ConceptSkin as Landscape
“We did the reading. You get the glow.”
ConceptPaper Plane Payday
“Get paid without the chase.”
ConceptThe Phone Siesta
“Nowhere to be. Beautifully.”
ConceptSunset on Tap
“Zero proof. All sunset.”
ConceptChef's Table for Dogs
“Dinner is served. Seriously.”
ConceptCoach Snail
“The slowest gym in town.”
Recast
One performance. Eight presenters.
The first tile is the original generated take for Casa Nowhere. Every other tile keeps that exact performance, with the same gestures, timing and handheld camera, and moves it onto another brand's presenter and setting with Genjutsu motion transfer. The lip movement still follows the original script, so the grid plays muted.
7 recasts of one take · muted
Hook A/B tests
Same brand. Different first second.
Every campaign ships with prepared test cells. The control keeps the original UGC hook; each variant changes one thing, such as the opening line or the presenter, and states what it would test. These are hypotheses ready to run, not results.
The Early Knot · Specialty café & bakery (local business). Each cell changes one thing against the control. Tap a video to play it with sound.
AControl
“3 things regulars order at 4 a.m.”
The original UGC hook, kept as the baseline the other cells are measured against.
BNew hook
“Regulars only: the pre-dawn order”
Tests whether a whispered insider-secret hook from the same baker earns more sends than the '3 things regulars order' list hook.
CNew presenter
“Save this for your sunrise walk”
Tests whether a young night-shift customer with a save-this discovery hook drives more saves and visits than the baker persona.
DVertical spot
“3 signs it's still warm”
Tests whether a people-free sensory proof list at the window beats the baker-led 15 s mini-list on completion and saves.
Lopeline · Running shoe DTC e-commerce. Each cell changes one thing against the control. Tap a video to play it with sound.
AControl
“Mile 1 vs mile 8 in rockers”
The original UGC hook, kept as the baseline the other cells are measured against.
BNew hook
“Flat foam is overrated”
Tests whether a myth/controversy hook from the same runner drives more comments and shares than the mile-1-vs-mile-8 comparison hook.
CNew presenter
“Couch to 10K: one change”
Tests whether a late-forties beginner with a transformation hook converts first-time buyers better than the fit twenty-something runner.
DVertical spot
“3 tests on the lime line”
Tests whether a people-free 12 s shoe-proof list beats the runner-POV 15 s cut on rewatches and saves.
Dew Diligence · Skincare serum DTC. Each cell changes one thing against the control. Tap a video to play it with sound.
AControl
“The niacinamide mistake I made for years”
The original UGC hook, kept as the baseline the other cells are measured against.
BNew hook
“Step 1: what it actually does”
Tests whether a how-to routine-step hook from the same presenter earns more saves than the confessional 'mistake I made' hook of the control.
CNew presenter
“Unfiltered skin at 38, part 1”
Tests whether a younger male skincare-newcomer with an unfiltered series hook widens reach and saves beyond the 50+ female persona.
DVertical spot
“3 ingredients, zero fragrance”
Tests whether a product-only ingredient list spot with no face or skin beats the presenter-led 15 s tip teaser on saves and completion.
Owed to Joy · B2B SaaS — invoicing and cash-flow app for freelancers. Each cell changes one thing against the control. Tap a video to play it with sound.
AControl
“Freelancers: the invoice thing nobody told me”
The original UGC hook, kept as the baseline the other cells are measured against.
BNew hook
“Stop chasing late payments: 3 steps”
Tests whether a numbered how-to hook from the same freelancer earns more saves than the 'thing nobody told me' insight hook.
CNew presenter
“3 months unpaid taught me this”
Tests whether a seasoned fifty-something consultant telling a pain story builds more trust and sign-up intent than the younger designer persona.
Casa Nowhere · Boutique hotel / travel. Each cell changes one thing against the control. Tap a video to play it with sound.
AControl
“Underrated hotel rule: phones take a siesta”
The original UGC hook, kept as the baseline the other cells are measured against.
BNew hook
“This hotel is airplane mode”
Tests whether a surprising-comparison hook ('hotel as airplane mode') from the same host earns more saves than the rule-reveal hook of the control UGC.
CNew presenter
“Hidden coves, part 1”
Tests whether a younger solo-guest creator with a series-plus-offer hook drives more shares and bookings intent than the older host persona.
Proof Negative · Non-alcoholic craft drink. Each cell changes one thing against the control. Tap a video to play it with sound.
AControl
“What I drink instead of a spritz”
The original UGC hook, kept as the baseline the other cells are measured against.
BNew hook
“3-ingredient mocktail that tastes like sunset”
Tests whether a recipe how-to hook from the same creator earns more saves than the 'what I drink instead' swap hook.
CNew presenter
“Ordering zero-proof at a rooftop like…”
Tests whether a late-thirties male humor persona with a comedic-POV hook widens reach beyond the early-twenties female creator.
DVertical spot
“3 reasons it tastes like sunset”
Tests whether a people-free flavor-list spot built on the sunset-can gag beats the presenter-led 15 s recipe cut on saves and loops.
Maître Dog · Premium pet food DTC. Each cell changes one thing against the control. Tap a video to play it with sound.
AControl
“POV: you switched your dog's food”
The original UGC hook, kept as the baseline the other cells are measured against.
BNew hook
“My mutt when the cloche lifts”
Tests whether a pet-reaction hook narrated by the same owner beats the 'POV: you switched food' hook on saves and rewatches.
CNew presenter
“Toughest critic rates dinner”
Tests whether an older female critic persona with a rating-review hook broadens appeal versus the younger male maître d' persona.
DVertical spot
“Sniff, stare, lick: the verdict”
Tests whether a dog-only 12 s taste-test list with no human on screen beats the owner-led 15 s comedy beat on shares.
Escargo · Boutique fitness studio. Each cell changes one thing against the control. Tap a video to play it with sound.
AControl
“3 reformer moves for desk-job hips”
The original UGC hook, kept as the baseline the other cells are measured against.
BNew hook
“Week 1 vs week 6, slow reformer”
Tests whether a before/after progress hook from the same coach beats the numbered-moves hook of the control on saves and profile taps.
CNew presenter
“First reformer class? Honest POV”
Tests whether a gym-shy first-timer in his late fifties converts beginners better than the expert coach persona.
DVertical spot
“Coach Snail's 3 rules of slow”
Tests whether a snail-only, people-free rules list in native vertical outperforms the coach-led 15 s follow-along preview on shares.
Behind the scenes
How it was made.
- Generation via the Higgsfield APIStills, edits, typography and motion were produced with a mix of image and video models, each picked for what it does best. Genjutsu motion transfer recast one filmed performance onto the other presenters.Marketing Studio ImageSoul 2Qwen Image 3 EditIdeogram 4Recraft 4.1Kling 3.0 ProSeedance 2.5Genjutsu
- Hooks with a reasonOpening lines and formats were informed by aggregate public TikTok engagement patterns, not by any single creator's content.
- Assembled with FFmpegVoice, captions, cuts and end cards were composed into finished ads with FFmpeg.
- Honest placeholdersEvery brand here is fictional. The landing pages carry no invented reviews, ratings or statistics; social-proof sections are deliberately left as waitlists and pledges.
Before you judge the demo
Questions about the demo.
Are these real brands or customer results?
No. All eight brands are fictional and were invented for this demo. Every image, voice, presenter and video is an AI-generated illustration. Nothing here is a customer project, and no campaign was run with a real audience.
Did the hook A/B tests run, and which variant won?
They have not run. Each brand has prepared test cells, a control and variants that change one thing, with the hypothesis each one would test. We publish no results, views or engagement numbers because none exist.
How were the hooks chosen?
From aggregate patterns in a bounded sample of public TikTok caption metadata, grouped by category: for example, how often list or how-to phrasing appears among posts with the highest share and save rates. No individual caption or creator video was reused.
Can I make a campaign like this in ViralCut Pro today?
Not as a self-serve feature yet. ViralCut Pro is preparing a done-for-you campaign service built on this pipeline. If a generated campaign would help your brand, email us.
Your footage next
Generated ads are the demo. Your recordings are the product.
ViralCut Pro turns the long videos you already have into captioned vertical shorts.