Strong inputs on Rhinobird reduce retries and credit waste. Generators read your product photo, category, and text fields literally—garbled photos or vague copy produce generic scenes and off-brand claims. This guide lists what to gather before you open AI E-commerce Image Sets, detail-page tools, or viral video, and how each required control behaves on the English UI.
Start with one hero product photo
Most commerce flows begin with Upload or Select Product. Treat this image as the single source of truth for shape, color, label, and packaging geometry.
- Lighting — even light, minimal color cast; avoid mixed tungsten and daylight that shifts SKU color.
- Background — plain or lightly textured backgrounds are fine; busy clutter confuses segmentation.
- Framing — show the full sellable unit (box, bottle, device) with edges not cropped unless the product truly is partial.
- Resolution — use the highest clean file you have; upscaling a tiny thumbnail cannot recover label text.
On viral video, Generate remains disabled until a product image is present. Plan the photo first, then choose Video Type and aspect ratio.
Category and naming
On AI E-commerce Image Sets, Category is required alongside the product image. Pick the category that matches how you merchandise the SKU—not a vague parent bucket—so scene templates and copy tone align with shopper expectations.
Product Name should match what customers see on the listing title or packaging front panel. Consistent naming helps generated headlines stay on-brand and makes History searchable when you revisit runs later.
Description and AI Writing Assistant
The Description field carries facts the model should not invent: materials, size, key ingredients, compatibility, warranty, or legally sensitive claims you want repeated accurately. Write short bullet-friendly sentences even if the UI accepts paragraphs.
When AI Writing Assistant is available next to Description, use it to expand structured facts into fluent marketing copy—but feed it accurate bullets first. The assistant amplifies what you provide; it does not replace legal review for regulated categories.
Skip empty superlatives (“best ever”) in favor of measurable benefits (“fits 15–17 inch laptops”, “SPF 50”, “USB-C charging”). Models treat concrete specs as anchors for USP frames and efficacy-style images.
Country / Language and aspect ratio
Country/Language steers spelling, idioms, and regulatory tone for generated text and on-image copy. Set it to the storefront or ad account locale before you generate, especially if you run separate US and UK listings.
Image Aspect Ratio on still tools should match the channel you will upload to—marketplace main images often differ from Instagram feeds. Changing ratio after a batch means regenerating, not cropping inside Rhinobird alone.
On video, pick Aspect Ratio among 9:16, 1:1, and 16:9 before generation so composition matches TikTok, Meta feeds, or YouTube pre-roll placements.
Planning counts on image sets
AI E-commerce Image Sets exposes count controls for sections such as Product Showcase, Key Selling Points (USP), Lifestyle & Usage, and Efficacy & Benefits. Decide how many frames you truly need per section before you press Generate—each increment affects credit cost shown on the button.
A practical first batch: enough showcase frames for carousel coverage, one or two USP frames for ad tests, and lifestyle frames only if your listing lacks real-world context. You can run a second batch with higher lifestyle counts after you validate the hero look.
Video-specific preparation
Beyond the product photo, choose Video Type intentionally:
- UGC Seeding — casual creator tone; script hooks should sound spoken, not brochure copy.
- TVC — polished brand voice; keep product name pronunciation-friendly in Description.
- Unboxing & Haul — mention pack contents in Description so hands interact with the right items.
- Pain Point Resolution — state the pain and resolution explicitly in Description so the narrative arc matches your claim.
If the product is reflective or transparent, spend extra care on the source photo—video models extrapolate glare and label curvature from that still.
Common failure states and fixes
- Generate stays gray — confirm product image uploaded and required fields like Category are set; sign in if prompted.
- Wrong product silhouette — retake or export a cleaner cutout; busy hands or props in the hero photo often bleed into generated scenes.
- Off-language copy — revisit Country/Language and regenerate; do not manually patch dozens of images if one setting fix solves the batch.
- Text hallucinations on labels — simplify the camera angle so the label is readable; extreme perspective labels may render gibberish.
Checklist before you spend credits
- One approved product photo uploaded or selected.
- Category and Product Name aligned with the live listing.
- Description with verified specs; optional pass through AI Writing Assistant.
- Country/Language and aspect ratio matched to destination channel.
- Counts and video type chosen deliberately; credit cost on Generate reviewed.
With assets prepared, move to tool selection or first ad creation articles to pick the exact generator route for your campaign.