Verified against Nano Banana / Gemini 3.1 Flash Image · 2026-07-31
Generate consistent alternate angles of the same product for a gallery
Produces a set of additional camera angles of one specific product from a single reference photo, holding the product's exact appearance fixed while only the viewing angle changes — for listings that need a multi-image gallery from one original shoot.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are generating additional camera angles of one specific product for a listing image gallery, using the uploaded reference photo as the ground truth for exactly what this product looks like. BASE PRODUCT PHOTO a wireless charging stand in matte white plastic with a rose-gold accent ring, photographed from a three-quarter front angle. This is the single source of truth for the product's actual appearance — every angle you generate must show the same product with identical color, proportions, material, and any visible branding or texture detail. TARGET ANGLES Generate these additional views: straight-on front view, direct top-down view showing the charging surface, and a rear three-quarter view showing the cable port. For each one, reason through what this product would actually look like from that specific angle given its true shape and proportions as shown in the base photo — don't approximate an angle by simply rotating a flat 2D image, actually reconstruct what a new physical vantage point would reveal, including features visible from the new angle that weren't visible in the original. CONSISTENCY REQUIREMENTS the rose-gold accent ring's exact width and position, and the exact placement of the small status LED near the base. Every generated angle needs to look like it came from the same photoshoot as the base photo and as every other angle in this set — if you find yourself rendering a detail differently in one angle than how it would logically follow from the base photo (a logo's exact position, a seam's exact path, a color that shifts warmer or cooler between angles), correct it before finalizing rather than delivering an inconsistent set. BACKGROUND AND LIGHTING MATCH same seamless soft-grey background and same overhead-softbox-with-side-fill lighting as the base photo. Keep the background treatment and the lighting setup's direction and quality identical across every angle in the set — a gallery where the lighting direction visibly shifts from one image to the next reads as several different, uncoordinated shoots rather than one coherent set, even if each individual image looks fine on its own. USE CASE a five-image gallery for an online marketplace listing — front, top-down, rear, and two three-quarter angles — keep this in mind for how many angles are actually useful; a small accessory might only need three angles to fully convey its shape, while a more complex product with functional details on multiple sides may genuinely need more to do its job in a listing gallery. SELF-CHECK BEFORE DELIVERING Before finalizing the set, compare every new angle back against the base photo's specific details — exact label text, exact proportions, exact color — and regenerate any angle where something has drifted rather than delivering a set with an inconsistency between images. OUTPUT The full set of requested angles, each as its own image, all traceable back to the same base product with no visible inconsistency between them.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
The instruction to reason through what a genuinely new vantage point would reveal — rather than approximate a new angle by rotating the existing flat image — targets the core technical challenge this task actually poses: a top-down view of a product photographed only from a three-quarter front angle shows real surface detail (a charging pad's texture, a previously hidden port) that simply doesn't exist anywhere in the original 2D photo, so generating that view correctly requires the model to infer the product's real 3D form from the one photo it has and extrapolate, not transform pixels that were never captured, which is a meaningfully harder and more error-prone task than a same-angle edit, and naming it explicitly sets the right expectation for how the model should approach it. Second, the self-check-before-delivering step matters specifically because multi-image consistency errors in a generated set are asymmetric in how they're noticed: a single product photo with a slightly-off detail might pass unnoticed, but a five-image gallery where the accent ring is a hair wider in one angle than another becomes obvious the moment a shopper scrolls through the set, precisely because the comparison is trivial to make when the images sit right next to each other — which is exactly the situation a single-image product shot never has to survive. Third, requiring identical background and lighting direction across the whole set, rather than treating each angle as its own independent generation, is what makes the output read as one coherent photoshoot instead of several separate renders that each happen to feature a similar product — a listing gallery's credibility partly rests on looking like it came from one real photo session, and a lighting direction that flips from one image to the next is a fast, easy-to-spot signal to a buyer that something about the listing's photos doesn't add up, even if they can't articulate the specific inconsistency.
Verified against
Nano Banana / Gemini 3.1 Flash Image Gemini 3.1 Flash Image · 2026-07-31
Changelog
- 2026-07-31 — Initial publish, verified against Nano Banana (Gemini 3.1 Flash Image) generating a five-angle gallery from one base product photo.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
