Google Maps Reviews Scraper 📍⭐

Extract Google Maps reviews fast, at scale and cheaply. No browser: the scraper talks to Google's own JSON endpoints, gets up to 60 reviews per request and never needs more than ~170 MB of memory.

• ⚡ Fast and light: runs in just 256 MB — 7,200 reviews from 100 places in under 2 minutes; thousands of reviews of a single place in minutes.
• 💸 Cheap: $0.25 per 1,000 reviews, and you only pay for reviews you get.
• 🧠 Deep: walks the complete review list of a place (tested on places with 45,000+ reviews). Optional Deep Search merges every sort order for the maximum possible yield.
• 🌍 Multilingual: original text plus Google's translation into the language you choose.

What data do you get?

Review	Reviewer	Place
Text and Google translation, original language	Name, ID, profile URL, photo	Title, Place ID, CID, feature ID
Stars, sub-ratings (Food, Service, Rooms…)	Number of reviews and photos	Category, address, city, country
Exact publication date + relative date	Local Guide status	Coordinates
Owner response (text + estimated date)		Rating, total reviews, rating distribution
Photos attached to the review		Google Maps URL, cover image
Review context (meal type, wait time, trip type…)		Permanently / temporarily closed

likesCount is always null: Google does not expose helpful votes to signed-out clients anymore.

Example item

{
  "reviewId": "Ci9DQUlRQUNvZENodHljRjlvT2toU05GTXRMVVp1WWpCcFEwaGlRbFY0YWpWTWNHYxAB",
  "stars": 4,
  "text": "量もあって美味しかったです！…",
  "textTranslated": "The portion was generous and delicious! …",
  "originalLanguage": "ja",
  "publishedAtDate": "2026-10-02T06:43:23.917Z",
  "publishAt": "4 days ago",
  "reviewDetailedRating": { "Food": 5, "Service": 2, "Atmosphere": 4 },
  "reviewContext": { "Meal type": "Dinner", "Wait time": "No wait" },
  "reviewImageUrls": ["https://lh3.googleusercontent.com/grass-cs/…"],
  "responseFromOwnerText": "Thank you for your review…",
  "responseFromOwnerDate": "2026-10-06T00:27:00.176Z",
  "name": "んご。",
  "reviewerNumberOfReviews": 29,
  "isLocalGuide": true,
  "title": "Gallagher’s Steakhouse NYC",
  "placeId": "ChIJFQ55zVdYwokRUyAEIvUGfHw",
  "address": "228 W 52nd St, New York, NY 10019",
  "totalScore": 4.6,
  "reviewsCount": 10008
}

How to use it

1. Add Google Maps place URLs (or Place IDs).
2. Set Number of reviews (leave the default to get all of them).
3. Pick a sort order — Newest is recommended.
4. Click Start and download the results as JSON, CSV, Excel, XML or HTML.

Supported inputs

• Place links: https://www.google.com/maps/place/…/data=…!1s0x…:0x…
• Review links: https://www.google.com/maps/reviews/data=…
• CID links: https://maps.google.com/?cid=12441467956046586299
• Place IDs: ChIJVVVVVVXlUVMRu-GPNDD5qKw, or links with query_place_id=
• Search links: https://www.google.com/maps/search/Museo+del+Oro+Bogota (the top result is used)
• Short links: https://maps.app.goo.gl/…

Input

Field	Description	Default
startUrls	Google Maps URLs (see above).
placeIds	Google Place IDs.
maxReviews	Max reviews per place.	10000000 (all)
reviewsSort	newest, mostRelevant, highestRanking, lowestRanking.	newest
deepSearch	When a sort order runs out first, continue with the other orders and merge (no duplicates).	false
reviewsStartDate	Only newer reviews: 2024-05-03 or 3 months, 8 days… With Newest the run stops as soon as older reviews appear.
language	Output language; foreign-language reviews get textTranslated.	en
reviewsOrigin	all or google (drop partner reviews, e.g. hotel booking sites).	all
personalData	Include reviewer name, ID, URL and photo.	true
proxyConfiguration	Fallback only — used while Google throttles direct connections.	Apify Proxy
maxConcurrency	Places scraped in parallel. Empty = automatic.	auto

Performance tips

• Memory: runs use 256 MB (the maximum), which keeps platform usage costs minimal. Expect roughly 70–150 reviews per second.
• One huge place is paginated sequentially (each page needs the previous one's token): ~60 reviews every ~0.3 s.
• Spending limit: the run stops by itself when your maximum charge for the run is reached, so you never pay for reviews that are not stored.

How it works

Google Maps now shows signed-out visitors a limited view (a handful of reviews, no pagination) and protects its own review endpoints with BotGuard tokens, which is why browser-based scrapers started returning 5 reviews or timing out. This actor reads reviews from the endpoint behind Google Search's reviews dialog, which still serves the complete, paginated list over plain HTTP, and enriches each review with the place details from Google Maps.

Use cases

Sentiment analysis, competitor monitoring, market research, reputation management, lead generation and collecting testimonials.

Development

npm install
npm test            # parser regression tests (fixtures in test/fixtures)
npm run start:local # runs input.example.json, results in ./storage