Pinterest holiday 2026: carousel stills from one product photo
Pinterest lists Carousel ads, Gift Badge and visual search for holiday 2026. Make four 2:3 stills from one product photo with the Sume Image API.

For a holiday Carousel, make four 2:3 stills from one product photo: a clean pack shot, a lifestyle scene, a detail close-up and a gift-ready scene. Pinterest's holiday page says visual searches grew three times faster than text search last holiday season, so the pictures carry the work. The Sume Image API can produce all four in one call with n: 4, and you review them before spending anything on video.
What does Pinterest say about holiday 2026?
The page gives no image sizes for Carousel ads. Check Pinterest's ad specs before final export.
| Topic | What the page says |
|---|---|
| Decisions | 68% say Pinterest helps them make decisions at every stage of their holiday shopping journey |
| Visual search | Visual searches grew 3x faster than text search last holiday season |
| Ad formats named | Carousel ads, Quiz Pin Extensions, the holiday Gift Badge program, creator content |
How do I make four cards from one photo?
Call POST /v1/images with your photo in input_references and aspect_ratio: "2:3". Read the model's supported_parameters first (GET /v1/images/models), because a model accepts only the ratios its catalog lists. Use aspect_ratio: "auto" only when you want the result to match the reference shape.
n runs from 1 up to the model's ceiling, which is 4 on Nano Banana 2.1 (read the n descriptor). Each image is billed as a full generation; failed generations are not billed. Results come back as signed Sume URLs (data[].url). Slow settings can return 202, and then you read the job result instead.
{
"model": "google/nano-banana-2.1",
"prompt": "Four-card holiday set for this candle: clean pack shot, cosy mantel scene, wax and wick close-up, gift-wrapped on a table. Keep the label exactly as in the reference.",
"input_references": [{ "type": "image_url", "image_url": { "url": "https://cdn.example.com/candle.jpg" } }],
"aspect_ratio": "2:3",
"n": 4
}What should I check before upload?
Check labels and text. Image models can alter small print, so compare each card with the original photo at full size. Keep one subject per card and put the offer in the Pin text, not in the picture, so you can change it without a re-render.
If a card should move, animate that one still with the video API using frame_images. The image step is the cheap place to find out what works.
What should I keep consistent across the stills?
A carousel reads as one set when the cards share a few things.
- The same product colour on every card.
- One background family, such as cream or deep green.
- The same crop ratio on all cards.
- A short caption that does not repeat the title.
Sources
Related posts
More in Use cases
- Podcast Video to Three Vertical Clips With Captions: $0.66
Cut three 45-second clips from a 30-minute video podcast, burn captions on each and pay $0.66 on Sume: video trim, standalone captions, and the limits to know.
- Podcast trailer: redo one part with previous_run_id, not a full rerun
Fix one weak scene in a podcast trailer by continuing the Format run with previous_run_id, so the agent keeps the rest. Rules, refusals and cost.
- Printable 2x8 inch bookmark is 1:4: cost of 6 to 24 designs
A 2x8 in bookmark is 1:4, or 600x2400 px at 300 dpi. Generate it directly with Nano Banana 2.1 on Sume: $0.15 at 2K, $0.20 at 4K per design. Costs for 1, 6, 24.
- Product launch teaser: one Format call with a $25 spend cap
How to produce a launch teaser through the Sume Format API: the call, the per-run cap, the idempotency key, and what the receipt shows when it finishes.
Written by Sume