Match Your Product Ad Image to the Shopping Moment: Season, Kitchen, Beauty & 9 Scenes

Ecommerce ad image scenes finally have a picker that matches how sellers actually think. When a homewares brand needs a Q4 ad, they think "I want a cozy autumn scene," not "template ID 7." When a skincare brand needs a Meta thumbnail, they think "marble vanity, beauty aesthetic," not "template with template_id = minimal-product." AmazVid Ad Image just shipped 20 templates across 9 scene tags with a secondary filter row that answers that vocabulary directly.

This is a small UX change that unlocks a big workflow: filter → preview → upload → generate in four clicks, credit deducted at 1 per image.

Start here: Ad Image gallery · Register free · Pricing · Ad images from listing photos.

The old gallery had a scaling problem

AmazVid Ad Image launched with 6 templates — small enough to skim. That grew to 10, then to 20 after this session's content pass. At 20 cards, a flat gallery becomes noise:

• A skincare brand scrolls past 6 seasonal templates and 4 fitness ones to find 2 beauty templates.
• A kitchen-appliance seller has to read every caption to spot the 2 kitchen scenes.
• A "just give me a product-only shot" merchant has to mentally filter out the 7 templates featuring human models.

Scanning cost > selection cost. Which means sellers pick the first plausible template and settle — losing the fit that a proper filter would surface.

The fix: scene tags + no-model cross-filter

Every ad_image template now carries two structured fields:

• scene — one of 9 tags: `season`, `fashion`, `beauty`, `kitchen`, `lifestyle`, `food`, `fitness`, `tech`, `promo`
• hasModel — boolean flag: true if a human is in frame, false if pure product still

The Template Gallery renders a second row of pill chips when the main Ad Image tab is active. Chips show: `All (20) · No model (13) · Season (5) · Fashion (2) · Beauty (2) · Kitchen (2) · Lifestyle (2) · Food (2) · Fitness (2) · Tech (1) · Promo (2)`.

Each chip prints its match count. Empty chips (a scene with zero templates) auto-hide. Switching the top-level tab away from Ad Image resets the filter to All so the next session starts clean.

The 20 templates + their scene tags

Season (5)

• `ai-holiday-glow` — hands cradling a mug beside amber string lights (no model)
• `ai-spring-bloom` — product among cherry blossoms with pastel morning light (no model)
• `ai-summer-beach` — product on golden sand with sea shells (no model)
• `ai-autumn-desk` — product on rustic desk with autumn leaves and latte (no model)
• `ai-winter-frost` — product on snow-dusted pine with cool blue light (no model)

Beauty (2)

• `ai-bold-hook-square` — 30-year-old woman with bold red lipstick on cyan gradient (with model)
• `ai-beauty-marble` — skincare product on blush marble vanity with dried florals (no model)

Kitchen (2)

• `ai-kitchen-marble` — product on white Carrara marble with morning window light (no model)
• `ai-kitchen-baking` — product on flour-dusted wood with rolling pin, warm oven glow (no model)

Fashion (2)

• `ai-minimal-luxury` — 40-year-old man in charcoal suit adjusting silver watch (with model)
• `ai-magazine-cover` — 24-year-old model in ivory blazer, Vogue-style hero (with model)

Food (2)

• `ai-ugc-holdup` — man in car holding snack bar UGC-style (with model)
• `ai-food-flatlay` — overhead flatlay with poached eggs, avocado, dill (no model)

Fitness (2)

• `ai-outdoor-lifestyle` — 32-year-old hiker at golden hour holding product (with model)
• `ai-fitness-gym` — product on matte black rubber gym floor with dumbbell (no model)

Lifestyle (2)

• `ai-lifestyle-card-story` — 30-year-old woman on sunlit yoga mat with water bottle (with model)
• `ai-flatlay-hero` — overhead flatlay on sand-beige linen with soft window light (no model)

Promo (2)

• `ai-sale-burst-square` — 25-year-old woman with orange retail bag on red-to-orange gradient (with model)
• `ai-neon-splash` — product on electric magenta-to-lime gradient with liquid splash (no model)

Tech (1)

• `ai-tech-workstation` — product on matte-oak desk with silver laptop, brass task lamp (no model)

Distribution:

• 13 no-model vs 7 with-model — most requests are "just the product," so the split leans that way.
• 5 season templates — seasonal ad rotation is the highest-frequency use case.
• Balanced coverage across the other 6 categories (2 each) plus a single Tech entry.

Two clicks to your ad

Old flow:

1. Open Template Gallery
2. Scroll all 20 cards
3. Read every caption
4. Pick the first plausible one
5. Click, modal opens
6. Upload SKU
7. Generate

New flow:

1. Open Template Gallery, click Ad Image tab
2. Click scene chip (Kitchen, No model, Season, etc.)
3. Click a template (now 1–5 options instead of 20)
4. Upload SKU in modal
5. Generate

Selection time drops from ~30 seconds of scanning to ~5 seconds of chip-clicking. Multiplied across a catalogue-refresh session that generates 50 ad images, that is 20 minutes saved per session.

When the No-Model filter matters most

The `hasModel` flag exists because sellers of certain product categories are wasting attention on model-led templates:

• Kitchen appliances — the product IS the hero. A model in frame competes with the kettle.
• Consumer electronics — same problem: the gadget needs to be the centerpiece.
• Home goods — furniture, decor, storage; the room styling matters more than a person.
• Food & beverage packaging — flatlays and countertop stills convert better than lifestyle model shots for the front-of-shelf image.
• Pet products — pet in frame yes, human model no.

For these categories, the "No model (13)" chip surfaces exactly the right subset. Two clicks, 13 hero-still options, upload, generate.

When you WANT a model in frame

• Fashion / apparel accessories — obvious.
• Beauty & cosmetics — bold-hook portrait converts better than a flatlay for lipstick / mascara.
• UGC-style social ads — hand-holding a product in a car / kitchen / gym reads more native than a studio still.
• Lifestyle framing — showing use-in-context, e.g. yoga water bottle on a mat.

For those, filter by scene (Beauty, Fashion, Lifestyle, Food) with hasModel implicit — the model-led templates dominate those buckets.

Under the hood — how filters work

Every template in `lib/templates/creative-templates.ts` now carries:

``` { id: 'ai-kitchen-marble', category: 'ad_image', scene: 'kitchen', hasModel: false, wizardConfig: { creativeType: 'image', ... }, ... } ```

The gallery calls `templatesByAdImageFilter(filterId)` which returns the matching subset:

• `'all'` — every ad_image template
• `'no-model'` — filters where `hasModel === false`
• Any scene ID — filters where `scene === scene`

Nothing dynamic on the server; this is a client-side filter. Fast.

Credit math for a scene-driven catalogue refresh

If your Q4 rollout needs 10 ads across 5 categories, budget:

• 10 Ad Images × 1 credit = 10 credits (~$2.50 pool)
• Add 2 Ad Creative videos for hero placements: 2 × 5 = 10 credits (~$2.50)
• Total: 20 credits ≈ $5

Compare to studio photography (~$500/day for a small in-house set) or freelance product photographer ($150+ per hero image). AmazVid comes in 99% cheaper and delivers in the same afternoon. Details in cost savings vs studio.

Comparison to other AI ad-image tools

The scene tag is AmazVid-specific — most tools organise by aesthetic (minimal, bold, luxury) not by shopping context. Which is fine for graphic designers; sub-optimal for sellers who think in terms of "the kitchen scene" or "the summer scene."

The prompt-only escape hatch

If your product needs a scene the 9 tags don't cover (e.g. "surgical clean-room stainless," "sailing yacht deck," "musical instrument close-up"):

1. Use the Prompt tab in Step 1 to describe your scene directly.
2. In Step 2, pick a base Ad Image template as the aesthetic anchor.
3. Generate. Your prompt overrides scene defaults; MiniMax renders the specific setting.

That works because Ad Image mode has the same prompt-only path video modes do. Details in prompt-only ai product video.

Next steps

Open the Ad Image gallery and pick your scene filter. Renders are 2 credits each (½-credit pairs on paid plans).

Related reads: Prompt-only video · Inclusive fashion casting · Packshot to ad creative images · AI ad animation for Meta / TikTok · AI product video pricing · Ad images from listing photos · Wearable vs Product vs Ad.