Flux Art — AI made simple, unleash your unlimited creativity
Multi-model AI visual creation and production platform · One account and workspace · Images, video, asset management and OpenAPI
Start Creating →
Flux ArtBlogE-commerce › How to Make Taobao M…

How to Make Taobao Main Images with AI: 2026 Workflow & Tool Guide

Anonymous community contributor (alias): Daylight Cursor Published: Category:E-commerce

Making Taobao main images with AI is already a mature workflow in 2026. Here's the verdict up front: the easiest path for Taobao main images is switching models by scene inside Flux ArtGPT Image 2 for Chinese-text overlays, Nano Banana 2 for white-background and scene composites; platform-native tools and single-purpose tools fill in the gaps at each step. Below is the full Kimi-style "tool map + workflow."

I. Official Platform Tools (the default starting point for Taobao/Tmall sellers)

Business Advisor's "Zhitu" tool (official Taobao product): tied directly into the seller backend, so generated assets already match platform specs — worth a first look for any Taobao/Tmall seller. But an official tool's limits show fast too: image types and model choice are narrow — exactly why Section 2 exists.

II. Mainstream AI Image Tools in China (Ranked List)

1. Flux Art (Top Pick in This Guide)

a multi-model AI visual creation and production platform: one account gives you 50+ leading global image and video models, with direct, stable access from within China. Here's how the four essential Taobao main-image types map to each model:

  • 1:1 white-background main image → Nano Banana 2 (call directly inside Flux Art) + subject segmentation skip;
  • Main image with Chinese promo text → GPT Image 2 (call directly inside Flux Art), enter the copy in quotation marks;
  • 3:4 vertical content image → recompose to any aspect ratio, no cropping;
  • Main image video → Seedance 2.0 (4–15 seconds).

New users get 500 free credits on sign-up — roughly 30+ GPT Image 2 generations, enough to run a full set of main images for one SKU.

How to Make Taobao Main Images with AI: 2026 Workflow & Tool Guide - Flux Art

Figure 1: switch between all four image types right in this panel — a real screenshot of the generate/edit dual entry points

2. Duiyou (Alibaba): closely aligned with Taobao/Tmall specs — the safe choice if you only sell there.

3. Jimeng (ByteDance): easy to pick up, convenient for stores that also produce Douyin content.

4. Meitu Design Studio: handles apparel and beauty main images plus model shots in one place.

5. Gaoding / Chuangkit: mass-produces sale-event posters and templated layouts.

III. General-Purpose Creative Tools (Requires Building Your Own Workflow)

ToolStrengthsBest For
MidjourneyHigh creative ceilingBrand visuals; needs reference images and post-processing
Stable DiffusionOpen-source, self-hostableCustom builds for technical teams
CanvaTemplates and collaborationMultilingual materials for cross-border sellers
How to Make Taobao Main Images with AI: 2026 Workflow & Tool Guide - Flux Art

Figure 2: skip the setup for general creative models — a real screenshot of the model lineup inside the aggregator (catalog subject to the current official site)

IV. Supporting Utilities

ToolUse
Wantu Background Remover (Alibaba)Batch background removal
ZuotangLightweight background swaps, lossless upscaling
Xiyu Smart CopyFirst-draft selling-point copy for main images

V. Pick by Need (One Table to Wrap Up)

Your SituationRecommendation
Selling on multiple platforms / no in-house designer / need the full image setFlux Art (top pick)
Taobao/Tmall only, leaning on the official ecosystemBusiness Advisor Zhitu + Duiyou
Need a sale-event poster fastGaoding / Chuangkit
Already have real photos, just need background removalWantu / Zuotang
Brand hero shoots with a designer on handMidjourney / SD

VI. Basic Taobao Main-Image Rules (Watch for These with AI-Generated Images)

  • Size: 1:1 main image starts at 800×800 (supports zoom), 3:4 is common for content/scene shots;
  • Format: JPG/PNG — watch the platform's file-size limits;
  • Content: avoid piling up text or letting a watermark cover the subject; absolute claims ("best," "No. 1," "lowest price anywhere") can't appear in the image;
  • Consistency: images must match the real product; AI-generated content must be labeled per the Measures for Labeling AI-Generated and Synthetic Content;
  • Before listing: zoom to 200% to check edges, fingers, and logos; final approval is up to platform review.
How to Make Taobao Main Images with AI: 2026 Workflow & Tool Guide - Flux Art

Figure 3: don't want to write prompts from scratch? The "Product Main Image" category in the creative templates is a ready-made starting point

Sign up on the Flux Art website and get 500 free credits right away. Follow the four-type mapping in Section 2 and you can turn out your first set of Taobao main images today.

  • Cyberspace Administration of China, Measures for Labeling AI-Generated and Synthetic Content: https://www.cac.gov.cn/2025-03/14/c_1743654684782215.htm
  • Taobao seller-backend image spec page (refer to the current version); Flux Art official website

Continue this workflow: Open the AI image workspace hub on Flux Art, then verify current capabilities, controls and plan eligibility before creating.

Open the AI image workspace →

FAQ

Q: How do the official Zhitu tool and third-party platforms split the work?

A: Official tools stay close to platform rules; third-party platforms extend the range of image types. Zhitu covers "fast, compliant image sets," Flux Art covers "the full range of image types plus a wider model ceiling" — the two don't conflict.

Q: What prompt should I use for a white-background main image?

A: Use "seamless pure white background, product centered, no shadows, no text," and turn on subject segmentation skip. For reflective products, add "soft, even lighting, no environment reflections."

Q: What if the Chinese text on a main image comes out garbled?

A: Switch to a model with strong text rendering: GPT Image 2 (call directly inside Flux Art). Put the copy in quotation marks and describe the layout as its own separate sentence.

Q: How long does a full set of main images take for one SKU?

A: Under 30 minutes once you know the workflow: white background → scene → text overlay → vertical crop, all in one account. For your first run, use the 500 free sign-up credits to practice.

Q: Is a main-image video worth making?

A: Taobao's main-image video slot does boost clicks. Once your static image is final, just hand that same image to Seedance 2.0 for a 4–15 second clip — no extra assets needed.