netOrb Amazon Scraper

Scrape Amazon search / category pages into clean product records — and get a ready-to-use affiliate link for each.

Built because Amazon hides price/rating/image behind JS on product detail pages, but exposes them right in the search-results HTML (data-asin, .a-price, .a-icon-alt, .s-image). One pass over the search page gets everything.

Input

Field	Type	Default	Notes
searchUrls	array	—	Amazon search/category URLs. Marketplace (amazon.ca/.com) is read from the URL. Required.
minRating	string	4.5	Only return products at or above this rating. Unrated products are skipped.
maxItems	integer	80	Stop after this many qualifying products. You're billed per returned result.
maxPagesPerUrl	integer	5	Result pages to follow per start URL (Amazon caps search at ~7 pages).
affiliateTag	string	(empty)	Your own Amazon Associates tag (e.g. yourtag-20), appended as ?tag=... to each affiliateUrl. Leave empty for a plain product URL. Never use someone else's tag.
proxyConfiguration	object	RESIDENTIAL	Residential proxy is effectively required — Amazon blocks datacenter/free IPs, so without it most runs return nothing.

Tip: each URL must be a search or category page (.../s?k=...), not a product (/dp/...) page. Filter on Amazon first (e.g. "4 stars & up") and paste that URL — fewer pages to scan.

Output (one dataset item per product)

{
  "asin": "B0CMYXLX13",
  "title": "Yealink SIP-T34W WiFi IP Phone ...",
  "brand": "Yealink",
  "rating": 4.6,
  "reviewCount": 128,
  "price": 129.99,
  "currency": "CAD",
  "imageUrl": "https://m.media-amazon.com/images/I/....jpg",
  "productUrl": "https://www.amazon.ca/dp/B0CMYXLX13",
  "affiliateUrl": "https://www.amazon.ca/dp/B0CMYXLX13/?tag=yourtag-20"
}

Any of price, rating, reviewCount, imageUrl, brand, currency may be null when Amazon's search card omits them. brand is best-effort (first title token) and can be wrong. If no products match minRating, the run succeeds with an empty dataset and a warning; if Amazon blocks every request or you pass a non-search URL, the run fails with a clear message rather than silently returning nothing.

Notes / limitations

• Amazon actively blocks scrapers. Residential proxy + session rotation + retries are built in, but success is not guaranteed; expect occasional captcha retries. Use this in compliance with Amazon's terms and applicable law.
• Uses CheerioCrawler (no browser) to keep compute/cost low. If blocking gets bad, switch the Dockerfile to apify/actor-node-playwright-chrome and use PlaywrightCrawler.
• Search-card selectors can drift when Amazon changes its layout — update the selectors in src/main.js if extraction returns empty.