A scenario prompt for captioning a multi-shot product gallery (hero, detail, lifestyle, packaging) without repeating the same sentence six times or stuffing keywords into shots that don't need them.
Works with Claude / GPT920 uses★ 4.4
The prompt
media-product-photography-captions
You're captioning a six-image gallery for one product listing: a ceramic pour-over
coffee dripper. The shots are:
1. Front view on white background
2. Top-down view showing the internal ridges that control drip speed
3. Close-up of the unglazed base, which is the detail customers ask about most
(does it get too hot to hold)
4. Lifestyle shot: dripper in use on a kitchen counter, steam visible
5. Packaging shot: the retail box, shown closed
6. Size comparison: dripper next to a standard 12oz mug
Write a caption for each image. Rules:
- Images 1 and 5 are close to decorative in context (they repeat what the product
title already says), so keep those captions short or skip them if they'd just
restate the title.
- Images 2, 3, and 6 are answering a specific customer question (drip mechanism,
heat safety, actual size), so their captions should say the thing a shopper is
scanning for, not just describe what's visible.
- Image 4 is mood/lifestyle. Don't invent a scent or sound in the caption just
because steam is visible; describe only what's confirmable from the photo.
- Flag which two captions are similar enough that a search engine could read them
as duplicate content, and rewrite one of them to be distinct.
[replace bracketed product details above with your own product and shot list]
Product galleries fail when every caption reads like a variation of the same sentence: “our dripper, front view,” “our dripper, top view.” Treat each shot as answering a different customer question instead, and let purely cosmetic shots (plain background, closed packaging) go caption-light so the ones that matter stand out.