Cover (white-bg hero)
SKU-locked packshot on solid background, product ≥60% of frame — generated from your real artwork, never re-drawn. No text.
AI product photography for Shopee: a packaging-true 9-slot gallery for every SKU
Shopee allows up to 9 images per listing (plus description images and a video slot), and Mall listings require at least 3 non-duplicate professional images with the cover on a solid — preferably white — background, product filling 60% of the frame, and no watermarks, borders or text on the cover. That is a lot of disciplined imagery per SKU. Berrys generates the full set from one reference packshot per product, locked to your real packaging, so a 200-SKU catalog looks like one brand instead of one thousand accidents.
Reviewed October 2026 · Prepared by Berrys · Technical specifications & Canvas node workflows
Shopee listing image requirements (checked 22 Oct 2026)
| Parameter / Slot | Marketplace & Production Requirement |
|---|---|
| Images per SKU | Up to 9 (cover included); up to ~12 extra description images via rich description |
| Aspect & size | 1:1 square. Work at 1024×1024 (many guides quote 500–800px minimums); JPG/PNG ≤2 MB per image |
| Cover image | Solid background (white preferred) · product ≥60% of frame · no watermark, montage, border, text or graphics · full product in frame |
| Cover logo | Top-left corner only, ≤10% of image frame |
| Mall minimum | ≥3 non-duplicate, professionally taken images; different angle per image |
| Non-cover images | Backgrounds/props allowed; product + props ≥50% of frame; close-ups and cropped details allowed |
| Cover models | Not allowed except Fashion, Health & Beauty, Sports & Outdoor categories |
| Fashion size charts | Footwear/apparel sub-categories require a separate size-chart image (rolled out from mid-2025) |
Free Browser Tool
Run free audit →Check your cover photo against Shopee sellers rules
Test your packshot cover image for solid white background (#FFFFFF), product frame-fill percentage and 1:1 aspect ratio before uploading to Seller Centre.
Use all 9 slots: a Berrys gallery plan
45° angle
Same lighting recipe, rotated product view. A different angle per image is a Mall requirement, not a nice-to-have.
Macro detail
Cap threads, fabric weave, port finish — zoom reassures buyers before they message you.
Lifestyle in context
Bathroom shelf, desk, kitchen counter — scene branches never touch the packaging pixels.
Feature callout
Text-overlay infographic generated inside the canvas (allowed on slots 2–9 only, never the cover).
Variant grid
List node fans every colorway/size into one comparison image with correct per-SKU labels.
Scale reference
Product beside a common object (hand, phone, can) to pre-empt 'smaller than expected' returns.
What arrives
Packaging + accessories shot. Directly cuts 'not as pictured' disputes.
Authenticity/certification
Critical for beauty and food: label close-up showing the real panel (cropped from your artwork, never redrawn).
Where generic AI breaks Shopee listings
Label drift is a return, not a typo
Models redraw text on every generation — a warped letter on your serum label becomes a 'barang tak sama gambar' dispute. Anything that redraws packaging instead of preserving it is unusable on Shopee. Berrys feeds your real artwork into the SKU node so typography and logo placement come from your files.
The 60% cover rule punishes 'pretty' scenes
Search results are a grid of thumbnails. AI lifestyle renders tend to shrink the product into the context; a cover where the product fills under 60% loses clicks to competitors even with better styling. Berrys exports cover-cropped masters at frame-fill ratios you set once in the workflow.
Storefront-level consistency
Buyers judge sellers by the weakest image. Shoots months apart give different color temperature; single-prompt AI gives different style per SKU. A saved Berrys workflow applies one lighting/background recipe across the whole store, so new SKUs match old ones automatically.
How Berrys produces the set
Brand System node
Load your logo files, hex palette and packaging references once. Every downstream generation inherits them — this is the layer generic prompt tools don't have.
Product references per SKU
Each product enters the canvas as its own File node reference wired to the Brand System, so rose-red lipstick and cherry-red shampoo never inherit the wrong packaging when you batch a catalog.
Multi-model canvas
Image nodes run Flux 3, Seedream 5.0 and GPT Image 2.5 (choose per job — Seedream for text-heavy compositions, Flux 3 for studio lighting); the video slot connects Google Veo 3.1, Omni Flash, Kling 4.0 and Seedance 2.5 under one subscription.
Production workflows in Berrys Canvas
One upload → full 9-slot gallery
1. File Node: Upload master packshot under neutral 5600K lighting.
2. Brand System Node: Assign brand hex palette and solid white background preset.
3. Image Node (Nano Banana 2.1): Render 1024×1024 cover hero (≥60% product frame fill, no text/props).
4. Fan-out: Duplicate to 45° angle, cap/macro detail, and real in-context shelf scene.
Double-digit sale variants (9.9 / 11.11 / 12.12)
1. File Node: Connect approved packshot master.
2. Brand System Node: Load festive promotional campaign theme while leaving top header clearance for platform badges.
3. Image Node (Nano Banana 2.1): Generate feed decoration banners without touching the clean white PDP cover.
Fill the video slot from a still
1. Image Node: Connect approved packshot hero.
2. Video Node (Omni Flash / Veo 3.1): Generate 5-second 360° product turntable orbit.
3. Video Trim Node: Crop and compress to <60s under 2MB for the Shopee video slot.
Berrys Plugins · Model Context Protocol (MCP)
Create in chat. Keep it on your canvas.
Already working in an AI agent? Connect Berrys to Claude, Cursor, ChatGPT, Codex, or Hermes. Your agent plans the shopee sellers suite and runs the shots on your Berrys canvas, with saved Brand Memory and credit approval in chat.
Connector URL:
https://app.berrys.io/api/mcpRemote MCPDrop-in Agent Brief (Shopee sellers)
Using Berrys, generate a compliant 9-slot Shopee listing gallery from my product reference in ./assets: 1. File node: Load product packshot under clean neutral lighting. 2. Brand System node: Set solid white background. Product must fill ≥60% of frame with zero watermarks or text overlays. 3. Image node (Nano Banana 2.1): Generate 1024×1024 square cover master. 4. Gallery fan-out: Generate 45° angle view, packaging macro detail, and contextual lifestyle scene. 5. Compliance check: Verify cover image meets Shopee Mall guidelines. Estimate credit cost and request approval before generating.
Compliance & production boundaries
Across Shopee, TikTok Shop and Amazon, the one-line rule is the same: images must accurately depict the product a buyer receives, and main images lean heavily toward 'real' rather than 'rendered' representation. Treat generated assets as a professional tool, not a license to embellish — photorealistic output from your actual packaging, QA'd before upload, never inventing contents, sizes or certifications.
When NOT to use AI generation for this
- Before your final packaging artwork is locked — AI will faithfully lock in a placeholder
- Color-critical goods (fabric, paint, cosmetics) where a monitor-difference return is costly: generate the set, but proof final color against the physical item
- Any slot where your live Seller Centre rules conflict with generated content — rules shift by site and month; verify, don't guess
Frequently asked questions
Can AI-generated images be used on Shopee listings?
Shopee Mall requires non-duplicate 'professional' images that accurately represent the product. Photorealistic outputs generated from your real packshots can meet that bar the same way studio edits do — but final accuracy QA is your responsibility, and for products where any rendering could mislead (complex assemblies, color-critical goods), keep the physical product shot as your cover.
Will the AI get my packaging text wrong?
Generic tools will — that is the classic label-hallucination failure. Berrys references your actual artwork files through the Brand System node and per-SKU File node references, so label text, fonts and logo placement follow your files instead of being redrawn.
How many images should I actually upload?
Use all 9 slots plus the video. Each slot answers a pre-purchase question; listings that stop at 2–3 images lose the buyers who otherwise message you and don't wait.
What does it cost to start?
Berrys Canvas has a free tier with 50 credits, no card required. Paid plans scale with generation volume, and Berrys Studio handles done-for-you campaigns when launches need art direction.
Related tool comparisons & guides
In-Depth Guides & Research from Berrys Journal
Create on-brand assets without booking a studio
Start free on Berrys Canvas — 50 credits, no card required — or send Berrys Studio a brief and get finished campaign visuals with agency art direction.