Shop Review Insights & Reply Writer

Turn a pile of marketplace reviews into three things: what's going wrong, what to fix first, and replies you can paste today.

What you get

• Health check: review count, average rating, negative share, unreplied negative reviews
• Monthly trend that pinpoints when ratings started to slip
• Worst products first, so you know where to look
• Complaint themes ranked by impact (shipping speed, packaging damage, defects, size/fit, not as described, missing parts, service...) with representative quotes
• Concrete fixes labelled by who owns them: product, listing page, fulfilment or customer service
• What buyers love, ready to reuse as selling points on your listing
• Ready-to-post replies to negative reviews, written in each reviewer's own language

How to use

1. Export reviews from your seller center (CSV or XLSX), or paste them.
2. Ask: "Why did my ratings drop last month?"
3. Get the report and the reply drafts.

Columns (rating, review text, date, product, seller reply) are detected automatically from common header names in many languages.

Works with

Shopee, Lazada, Tokopedia, TikTok Shop, Amazon, momo, or any CSV/XLSX.

Languages

Reviews in Traditional/Simplified Chinese, English, Japanese, Korean, Vietnamese, Indonesian and Thai. Report in your language; replies in the reviewer's language.

Example prompts

• "幫我分析這份蝦皮評論匯出檔，為什麼 10 月負評變多？"
• "Which product has the worst reviews and what should I fix first?"
• "Write polite replies to all unreplied 1-2 star reviews."

Requirements

Runs locally with Python 3 (openpyxl only for .xlsx). No API key; your shop data never leaves your machine.