Free Google Shopping Reviews Scraper — Product Reviews + Rating Breakdown by SKU

A Google Shopping reviews scraper that returns the full product-review payload Google aggregates across retailers — average rating, total review count, the 5/4/3/2/1-star breakdown, and individual user reviews with author, source retailer, date, and full text — as clean structured JSON. Pass a Google Shopping sku (catalogid) or gpcid (product cluster id) plus a country code, and the actor paginates the reviews for you. $0.01 per product processed. No Google Merchant Center API account, no DataForSEO subscription, no Selenium.

What you get

• Product title as Google renders it on the cluster page (e.g. "Sony WH-1000XM5 Wireless Noise-Canceling Headphones")
• review_summary — average_rating (rounded to 1 decimal, the headline number Google shows), exact_average (the un-rounded float for tie-breakers), and total_reviews (integer count across all retailers)
• review_breakdown — the percentage distribution across 5, 4, 3, 2, and 1 stars (e.g. {"5": "78%", "4": "12%", "3": "5%", "2": "3%", "1": "2%"})
• user_reviews[] — individual reviews each with title, source (the retailer that hosted the review — Amazon, Best Buy, BHPhoto, Newegg, etc.), author name, rating as a string, posting date, and full review text
• Up to 10 reviews per page × 10 pages = 100 reviews per product per run
• The original sku and gpcid echoed back so you can join review records to your product catalog
• country echoed back — the same product cluster has different review pools per Google Shopping country edition

Why scrape Google Shopping reviews

Google Shopping is the only place on the web where reviews from Amazon, Best Buy, Walmart, Target, B&H Photo, Newegg, Adorama, John Lewis, MediaMarkt, Bol.com, and dozens of other retailers are aggregated under a single product cluster. For a brand manager, e-commerce analyst, market researcher, or product-launch team, this aggregated review pool is more valuable than any single retailer's reviews — it's a cross-retailer demand signal that tells you what customers actually say about a product no matter where they bought it. But Google does not expose this data through any public API. The official Merchant Center API only returns reviews for products you sell as a verified merchant; the Custom Search Engine doesn't index review text; and the only third-party APIs that return Google Shopping reviews (DataForSEO, SerpApi) charge $4-15 per 1000 product reviews with rate limits.

This actor solves that by hitting Google Shopping's internal /async/skp_rj endpoint directly with curl_cffi (Chrome TLS impersonation) — the same JSON endpoint Google's product page calls when you click "See more reviews". Pass a sku (the catalogid you'll see in any Google Shopping URL like shopping.google.com/product/123456) or a gpcid (the cluster id) plus a country code, and the actor builds the protobuf reviewSettings parameter, paginates through all pages, and returns the full payload for $0.01 per product.

Concrete buyer math: a brand manager monitoring 200 SKUs across 5 countries weekly pays 200 × 5 × $0.01 = $10 per weekly run — versus DataForSEO at $0.045 per product (the same 1000 calls would cost $45) or SerpApi at $0.005 per query × 5-10 paginated calls per product = $25-50. A market-research firm doing one-off competitive sweeps on 5,000 SKUs pays $50.

Input

Field	Default	Description
products	required	Array of {sku: string, country: string} or {gpcid: string, country: string} objects. sku = the Google Shopping catalogid; gpcid = the product cluster id. Mix both in one run
max_pages_per_product	3	Pages of reviews per product (10 reviews per page; 1-10 pages = 10-100 reviews)
concurrency	5	Parallel product fetches (1-20)
default_country	us	Used when a product entry doesn't specify country

When you only have a sku, the actor first hits the Google Shopping product page to resolve the corresponding gpcid, then builds the reviewSettings protobuf and starts paginating. When you already have the gpcid, the actor skips the resolution step and goes straight to reviews — faster and one fewer network call.

Output

{
  "sku": "8062340666125262933",
  "gpcid": "158194847738236757",
  "country": "nl",
  "title": "Sony WH-1000XM5 Wireless Noise-Canceling Headphones",
  "review_summary": {
    "average_rating": 4.6,
    "exact_average": 4.62,
    "total_reviews": 12847
  },
  "review_breakdown": {
    "5": "78%",
    "4": "12%",
    "3": "5%",
    "2": "3%",
    "1": "2%"
  },
  "user_reviews": [
    {
      "title": "Best noise cancelling I've ever owned",
      "source": "Amazon",
      "author": "Marcus W.",
      "rating": "5",
      "date": "2026-03-12",
      "text": "Replaced my XM4s and the difference is immediate — the new processor crushes airplane noise even better, and the call quality on the new mics is finally good enough for daily standups."
    },
    {
      "title": "Great sound, mediocre build",
      "source": "Best Buy",
      "author": "Linda K.",
      "rating": "4",
      "date": "2026-02-28",
      "text": "Audio is incredible, but the headband flex feels cheaper than the XM4. For the price I expected metal, not plastic."
    }
  ]
}

review_summary.exact_average is the un-rounded float Google uses internally for ranking — handy when you need to break ties between two products both displayed as "4.6 stars". source tells you which retailer hosted the original review, so you can weight reviews differently (e.g. trust Best Buy reviews higher than aggregator-site reviews).

Use cases

Brand manager monitoring product reception across markets. You're a product manager at a consumer-electronics brand. You ship 50 SKUs across the US, UK, DE, NL, and JP Google Shopping editions. Schedule the actor weekly with all 50 SKUs × 5 countries × max_pages=10 (100 reviews per product per market). 250 product-runs × $0.01 = $2.50 per weekly run — and you have country-segmented sentiment for every SKU you ship.

E-commerce category manager benchmarking competitors. You sell wireless headphones on a marketplace and want to know what customers actually complain about across the top 30 competitor models. Pull 30 gpcids × 100 reviews each = 3,000 reviews for $0.30. Run topic modeling on text and segment by rating. Within an hour you know that "comfort over long calls" is the #1 complaint cluster — your next product brief writes itself.

Market-research firm building a product-launch report. Your client wants a trend report on the air-fryer category for Q2 2026. Pull 200 air-fryer SKUs across the US and EU (1,000 product-runs = $10) and you have a 100,000-review corpus segmented by rating, retailer source, and date. Ship the report; pocket the difference between the $10 of scraping and the $5,000 retainer.

Consumer-insights team training NLP / sentiment models. You need 1M labeled (rating, text, source) tuples for fine-tuning a multi-retailer aspect-based sentiment model. Pull 10,000 SKUs × 100 reviews each = 1M reviews for $100. Each review carries the source retailer label, which becomes a "domain" feature in your model — invaluable when retailer-specific reviewers have different vocabulary norms.

How it compares

Service	Price per 1000 product calls	Cross-retailer aggregation	Star breakdown	Per-review source
This actor (s-r/free-google-shopping-reviews)	$10	yes (Amazon, Best Buy, Walmart, etc.)	yes (5/4/3/2/1 %)	yes
DataForSEO Google Shopping Reviews API (#2 on google shopping reviews api SERP)	~$45	yes	yes	yes
SerpApi Google Product API (#4 SERP, similar feature)	~$25-50 (paginated)	yes	partial	yes
Google Merchant API Reviews (#1 SERP)	free (verified merchants only)	own products only	yes	no
Bright Data Google Reviews API (#6 SERP)	per-request, similar tier	yes	yes	partial

The honest competitive position is: the only free or low-cost path to Google Shopping's cross-retailer aggregated reviews is scraping. The Google Merchant API is gated to product owners; everything else (DataForSEO, SerpApi, Bright Data, Outscraper) is a paid scraping wrapper. This actor is a direct, cheap version of the same wrapper.

Pricing

This actor uses Apify's pay-per-event monetization at $0.01 per product processed — that's $10 per 1000 products, with each product yielding the summary + breakdown + up to 100 paginated reviews depending on max_pages_per_product. All pricing is pay-per-event — you only pay for products that successfully resolve. No actor-start fee, no per-compute-unit charges, no per-review surcharge — pagination is included in the per-product price.

Limits and gotchas

• Maximum 100 reviews per product per run (10 pages × 10 reviews) — Google paginates beyond this but quality drops sharply (older, less-relevant reviews); for the rare product with >100 useful reviews, run multiple times with different country codes since each country edition has a different review pool
• The country field controls which Google Shopping edition's reviews pool you query — us returns mostly English Amazon / Best Buy reviews, nl returns mostly Dutch Bol.com / Coolblue reviews, de mostly MediaMarkt / Otto / Amazon DE. The aggregated total_reviews count is country-specific
• Some products are gated behind region locks (a US sku queried with country=jp may return zero reviews if the cluster doesn't exist in JP); the actor returns an empty user_reviews[] rather than erroring
• Apify residential proxy is not required — the cookie pool plus curl_cffi's Chrome TLS fingerprint is enough for sustained throughput
• Cold-start time is ~2-3 seconds for the first product (one extra fetch to resolve sku → gpcid); subsequent products in the same run reuse the session pool
• Reviewer date strings come back in Google's local format for the country ("March 12, 2026" for us, "12 maart 2026" for nl); normalize downstream if you need a single date format
• The sku you pass must be the Google Shopping catalogid (the long numeric ID in shopping.google.com/product/<catalogid>), not a retailer SKU like Amazon's ASIN

FAQ

How do I get a Google Shopping product SKU? Open any product page on Google Shopping (shopping.google.com/product/...) — the long numeric ID at the end of the URL is the catalogid. That's the sku you pass to this actor. Alternatively, if you already have the gpcid (cluster id) from a previous run, pass that and skip the resolution step.

What's the cost to scrape 1000 Google Shopping product reviews? 1000 × $0.01 = $10 for 1000 product reviews-runs (each yielding the summary + breakdown + up to 100 paginated reviews). For comparison: DataForSEO charges ~$45 for 1000 product calls; SerpApi charges ~$25-50 depending on pagination depth.

Can I scrape Google Shopping reviews without a Merchant Center account? Yes. The official Google Merchant Center API only returns reviews for products you sell as a verified merchant. This actor scrapes the public Google Shopping product cluster pages, which expose every review across all retailers. No Merchant Center, no Google Cloud account, no API key.

Why are the reviews different in different countries? Google Shopping runs separate product clusters per country edition — a gpcid for the same physical product (e.g. a Sony WH-1000XM5) has a different review pool in us (Amazon US, Best Buy, B&H), nl (Bol.com, Coolblue, MediaMarkt NL), and de (Amazon DE, MediaMarkt DE, Otto). Run the actor per country to get the full multi-market picture.

Will Google rate-limit my scraping? At default settings (concurrency 5, fresh session cookies) the actor sustains ~10 products/second. Push concurrency to 20 for batch jobs; for sustained >100/sec throughput, supply a residential proxy URL via the PROXY_URL env var.

Related Actors

• Free Google Maps Reviews Scraper — every review for any Google Maps place URL (local-business reviews, not product reviews)
• Free Google Maps Scraper — find local businesses by keyword + city with full contact data
• Free Google Hotels Scraper — Google Hotels search with per-room pricing, 18-31 OTA comparison, and review breakdown