Image Account Content Manager

Problem

Image-based product posts for WeChat official accounts and image accounts often get stuck on the same issues: plain product photos lack scene context, titles and copy need multiple angles, and tags plus product links still require manual assembly. For creators, this makes content production fragmented and hard to sustain.

How it works

The skill accepts product images or text descriptions. In image mode, a multimodal parser extracts color, packaging, brand, and product form, then converts the details into generation prompts. In text mode, it expands the selling point, use case, and target audience into content. The output is a fixed package:

• Scene images: 3-5 vertical 750x1000px images, covering gifting, personal use, unboxing, and lifestyle angles.
• Titles: 3 options, leaning emotional, pain-point, and curiosity-driven.
• Copy: 5 pieces, each ≤300 characters, easy to copy into image-account posts.
• Hashtags: 5 industry or trending tags.

Boundary: It is a text-to-image pipeline, not image-to-image. Product fidelity is around 80-90%, so brand fonts and fine textures may vary. Product links must be added manually through WeChat or Video Account shop permissions. It fits physical products such as food, beauty, mother-baby, home, and electronics.