Cross-Border E-Commerce Product Copy Generator

The Problem

Cross-border product copy often needs repeated rewrites for Amazon, Temu, Shopee, and TikTok Shop. Handwriting multilingual listings can miss SEO keywords, exceed character limits, hit platform restrictions, or mix too many variables in A/B tests. This skill constrains copy generation to a repeatable input set: product_name, platform, language, keywords, features, tone, ab_version, and ab_diff_type, then outputs structured product copy.

How It Works

It starts with product information and generates titles, short descriptions, detail descriptions, and bullet points according to platform style. It weaves supplied keywords naturally into the copy instead of appending them, and adjusts tone across professional, marketing, luxury, and casual modes. In A/B mode, it can produce 1–5 versions while controlling variation through keyword, tone, structure, or content, keeping core selling points and keywords consistent.

It also handles common failure points:

• Multilingual adaptation: supports en, zh, bilingual, es, fr, de, ja, and ar, with localized phrasing and layout.
• Compliance checks: flags infringing, prohibited, or platform-specific expressions, then replaces them or suggests fixes.
• Character adaptation: compresses or expands copy against platform limits and reports module character counts.
• Structured output: returns single or multi-version results as JSON, making it easier to feed into listing tools or test workflows.

Boundaries

This is useful when product name, selling points, and keywords are already clear. For regulated categories, brand claims, certifications, or medical-style statements, human review is still required. When running A/B tests, avoid stacking multiple variables; keep facts and keywords stable and change one dimension at a time.