Amazon Full Data Extractor

Problem

Amazon competitor data is often scattered across PDPs, search results, category pages, seller storefronts, Best Sellers lists, and review pages. When tracking prices, observing BSR movements, extracting complaint themes, or building competitor profiles, copying page data and re-parsing JSON by hand is brittle and hard to audit. The Amazon Full Data Extractor skill narrows this workflow to a set of MCP tools that return structured fields directly, rather than leaving raw page parsing to the model.

How it works

It covers six common Amazon data paths:
- ASIN detail: get_amazon_product returns title, price, star, rating, features[], bestSellersRankItems[], aiReviewsSummary, and PDP review snippets.
- Keyword search: search_amazon returns the first-page ASIN list with price, rating, monthly sales string sales, badges, and sponsored.
- Category / seller / rankings: list_category_products, list_seller_products, list_bestsellers, and list_new_releases handle category listings, storefronts, bestseller lists, and new-release lists.
- Bulk reviews: get_amazon_reviews fetches multiple pages of reviews for VOC, sentiment, and complaint clustering.
- Fallback filtering: when price range, sort order, or pagination cannot be expressed by dedicated parameters, scrape_url accepts a full URL.

The key step is choosing the right tool and extracting the correct field path. For example, bestseller/new-release recsList values are JSON strings that need a second parse; search_amazon expects keyword, not keywords; and concurrent MCP calls should stay around two to avoid 9200 errors.

Boundaries

It is Amazon-only and should not be used for Walmart, eBay, Shopify, or other marketplaces. Full GTM reports, niche filtering, and listing optimization belong to more specialized skills. Review scraping consumes credits per page, so page counts should be confirmed in advance; missing fields should be reported as missing rather than estimated.