Amazon Product Research + AI (Market & Price Finder)

Turn an Amazon search into a clean, structured product dataset plus an AI market read‑out — built for sellers, dropshippers and product researchers who want to size up a niche fast.

For each keyword the actor collects products from Amazon search (across pages) and returns, per product: ASIN, title, price, list price, rating, review count, image, product URL, Prime flag and whether it's sponsored. Then the built‑in AI analyzes the whole keyword's market:

• Price segments (budget / mid / premium bands)
• Competition level (low / medium / high)
• Opportunity gaps — under‑served angles and weak spots
• Best‑value and highest‑rated picks (by ASIN)
• A recommended price point to enter the market
• A positioning angle tailored to your product (optional)

All price statistics (min / max / average / median, average rating, total reviews) are computed in code and handed to the AI — so the AI explains the market without ever inventing numbers.

Input

Field	Type	Description
keywords	array (required)	Amazon search terms / categories, one per line.
domain	select	Amazon marketplace (amazon.com, amazon.co.uk, amazon.de, …). Proxy country + currency auto‑matched.
maxProductsPerKeyword	integer	Products to collect per keyword. Default 20, max 100.
aiAnalysis	boolean	AI market analysis. Default true. Off = raw extraction only (no AI charges).
yourProduct	string	What you sell, so the AI tailors positioning advice. Optional.
aiLanguage	string	Language for AI‑written fields. Default English.
proxyConfiguration	object	Residential proxy is required for Amazon.

Example input

{
  "keywords": ["wireless mouse", "yoga mat"],
  "domain": "amazon.com",
  "maxProductsPerKeyword": 30,
  "aiAnalysis": true,
  "yourProduct": "ergonomic vertical mouse for designers"
}

Output

The dataset holds two kinds of rows (distinguished by recordType):

Product rows (recordType: "product")

{
  "recordType": "product",
  "keyword": "wireless mouse",
  "position": 1,
  "asin": "B001DHECXA",
  "title": "TECKNET Wireless Mouse, 2.4G USB Computer Mouse...",
  "url": "https://www.amazon.com/dp/B001DHECXA",
  "priceText": "$9.99",
  "price": 9.99,
  "listPrice": null,
  "rating": 4.4,
  "reviewsCount": 49873,
  "image": "https://m.media-amazon.com/images/...jpg",
  "sponsored": false,
  "prime": true,
  "currency": "USD"
}

Market analysis row (recordType: "market_analysis", one per keyword)

{
  "recordType": "market_analysis",
  "keyword": "wireless mouse",
  "productCount": 30,
  "priceMin": 6.99,
  "priceMax": 49.99,
  "priceAvg": 18.4,
  "priceMedian": 16.99,
  "avgRating": 4.4,
  "totalReviews": 812340,
  "aiAnalysis": {
    "marketSummary": "...",
    "priceSegments": ["Budget < $12", "Mid $12–25", "Premium > $25"],
    "competitionLevel": "high",
    "opportunityGaps": ["Quiet-click models with long battery", "..."],
    "bestValueAsin": "B004YAVF8I",
    "highestRatedAsin": "B087Z5WDJ2",
    "recommendedPricePoint": "...",
    "positioningAngle": "..."
  }
}

Tip: filter the dataset by recordType to separate the product table from the market summaries.

How it works & limits

• Uses Apify Residential Proxy with the country pinned to the marketplace — this keeps the currency correct and the block rate low.
• Reads everything from Amazon search pages (one page ≈ 20+ products), which keeps cost low. It does not open every product detail page.
• Amazon caps organic search around 7 pages (~100+ results); that's the practical ceiling per keyword.
• Occasionally a page may be challenged; the actor retries with fresh proxy sessions and moves on.
• Respect Amazon's Terms and applicable laws when using the data.