What is Google Reviews Scraper?
Google Reviews Scraper reads the reviews of one place on Google Maps and returns them as records, one per review. You name the place, pick an order such as newest first or lowest rated, and get ratings and texts back in a form you can count, filter and quote.
It is the step after discovery. Google Maps Scraper finds the places in an area; this tool reads what customers wrote about the ones you picked.
You ask
Get the 20 newest Google reviews of The Metropolitan Museum of Art in New York with rating, date, text and visit details.
It runs
From your AI assistant over MCP or from your code over REST
You get
- Review
- Media and source
- Reviewer
- Visit details
- Place and topics
- Watch your own locationsPull the newest reviews every morning and catch a one-star review before it sits unanswered for a week.
- Read competitors' complaintsSort a rival's reviews by lowest rating and see which problems customers report again and again.
- See what people talk aboutReview topics come back with mention counts, so you learn that 532 reviews of The Met mention helpful staff without reading them all.
- Compare branches of a chainSend the same request for every store and line up ratings, wait times and recurring words.
- Feed an analysisHand the texts to Claude or ChatGPT in the same chat to group praise and complaints, or load them into your own sentiment model.
What data does the Google reviews scraper extract?
Each review is one record, and every response also describes the place. These are the fields our test run returned. A field stays empty when the reviewer left it out.
Review
- Star rating, 1 to 5
- Review text
- Text in its original language
- Date posted, relative and exact
- Date of the last edit
- Likes from other users
Media and source
- Photos attached to the review
- Where the review was posted
- Review ID
- Direct link to the review
Reviewer
- Name and profile link
- Local Guide badge
- How many reviews they have written
- How many photos they have posted
- Profile picture
Visit details
- Weekday or weekend visit
- Wait time
- Whether a reservation is recommended
Place and topics
- Place name, address and type
- Average rating and total reviews
- Review topics with mention counts
- Token for the next page
Visit details come only from reviewers who answered Google's follow-up questions: 11 of our 20 reviews had them. Reviewer names and profile links are personal data, so this page masks them and you should store them only when you need them.
Sample output: the 20 newest reviews of The Met
The table at the top of this page is a real run from 16 November 2026: the 20 newest Google reviews of The Metropolitan Museum of Art in New York, a place with a 4.8 average across 95,517 reviews. The request took under a second, and all 20 reviews had been posted within 25 hours of it. Below the chart is one review in full.
One result in full
One review in full
- Rating
- 5 of 5
- Posted
- 16 Nov 2026, 01:16 UTC (shown on Google as “15 hours ago”)
- Reviewer
- M•••• G., Local Guide, 175 reviews and 693 photos on Google
- Text
- Had a fantastic experience at The Met! Getting in was super easy with mobile tickets, and the museum was surprisingly easy to navigate and find the exhibits we wanted to see. Even on a weekend, it wasn’t…
- Visited on
- Weekend
- Wait time
- No wait
- Reservation recommended
- Yes
- Photos attached
- 10
- Likes
- 0
- Source
- Review ID and link
- Returned with every review
What this sample tells you
19of 20
19 of 20 reviews give five stars and one gives four. A newest-first slice of a popular place says little about its problems, so sort by lowest rating when you look for complaints.
9of 20
Only 9 of 20 reviews have any text. The other 11 are a star rating alone, so count the texts before you plan an analysis on them.
5of 20
5 of 20 reviews carry photos, 29 in total, and 14 of 20 reviewers are Local Guides.
The run returned 10 review topics for the place. “Helpful staff” leads with 532 mentions, ahead of sculptures with 261 and the Egyptian section with 178.
How to scrape Google reviews with your AI assistant
Add the MCP server
Use Try it now to connect Claude, ChatGPT, Cursor or another MCP client. One sign-in with your AI/ML API account covers every tool.
Name the place
Give the business name and city, or paste a place from a Maps search. Your assistant looks the place up before it reads reviews.
Pick the slice
Say how many reviews you want and in which order: newest, most relevant, highest or lowest rating. Add a word to filter on, such as parking or refund.
Ask your real question
Have the assistant count ratings, group complaints by theme, quote examples or save the table as CSV.
Prompts to try
Use the AI/ML API MCP Google Reviews Scraper tool. Get the 50 newest Google reviews of The Metropolitan Museum of Art in New York and tell me how many mention crowds or queues.
Use the AI/ML API MCP Google Reviews Scraper tool. Find the lowest-rated Google reviews of the IKEA in Brooklyn that mention delivery and list the problems people describe.
Use the AI/ML API MCP Google Reviews Scraper tool. Compare 30 recent Google reviews of the British Museum and the Louvre: average rating, reported wait times and the three most common complaints.
Use the AI/ML API MCP Google Reviews Scraper tool. Show the review topics Google lists for Joe's Pizza on Carmine Street, New York, with the mentions for each, then pull 10 reviews for the top topic.
Settings for a reviews request
In a chat your assistant fills these in from your sentence. In code you send them yourself.
| Setting | What it controls | Tip |
|---|---|---|
| Place | Google place ID or Maps feature ID of one business | Copy it from a Maps search result, or let your assistant find it by name and city |
| Number of reviews | Up to 100 on an unfiltered first request, up to 20 on filtered requests and later pages | Start with 20 to see the mix, then page on |
| Order | Most relevant, newest, highest rating or lowest rating | Newest for monitoring, lowest for complaint research |
| Keyword | Only reviews that match a word or phrase | Use the word a customer would write, such as “refund” |
| Topic | Only reviews under one topic from the place's topic list | Use a topic or a keyword in one request, not both |
| Language and country | Language of labels and place details, country code for localization | Match the market you study |
| Next page | Continuation token from the previous response | Pass it back unchanged to get the next batch |
The place's average rating and total review count come back with every request, so you can see where a place stands without downloading all of its reviews.
What teams do with Google reviews data
Reputation monitoring
Pull the newest reviews of each of your locations on a schedule and alert the manager when a one- or two-star rating arrives.
RatingDateReview link
Complaint research
Sort a competitor's reviews by lowest rating, filter by a word such as “wait” or “refund” and list what goes wrong most often.
RatingTextKeyword filter
Location benchmarking
Request the same slice for every branch of a chain and compare averages, review counts and the topics customers raise.
Average ratingTopicsText
Visit and event planning
Read wait times and reservation answers before you send a group or a client to a museum, restaurant or venue.
Visited onWait timeReservation
Sentiment training data
Collect labelled examples, with the star rating as the label and the text as the input, for a classifier or an evaluation set.
RatingTextOriginal text
Briefs written by your agent
Ask your assistant to read 50 reviews, quote the clearest examples and write a one-page summary for a client or a landlord.
TextDateTopics
Google Maps reviews API for your code
In your code
REST · your AI/ML API key, JSON records back
The same tool runs over REST with your AI/ML API key. Send a place ID, the number of reviews and an order, and you get JSON with the review records, the place's average rating and total count, the topic list and a token for the next page.
A typical pipeline finds places with a Maps search, stores their place IDs and requests the newest reviews for each ID on a schedule, keeping review IDs so it never stores the same review twice. Use MCP when someone asks questions in a chat and REST when a job runs unattended. Both draw on one balance. Start with the Toolbox REST quickstart.
Other ways to get Google reviews
Google has official routes to reviews, but each covers only part of what a review study needs.
| Google Places API | Google Business Profile API | Your own scraper | ||
|---|---|---|---|---|
| Which reviews | Any public place, sorted, filtered and paged | Up to 5 reviews per place | All reviews, only for locations you manage | Whatever you can keep it collecting |
| Inside an AI chat | Yes. Over MCP | Only if you build it | Only if you build it | Only if you build it |
| Best for | Research on your places and your competitors | Showing a handful of reviews in an app | Replying to reviews of your own business | Teams with engineers to maintain it |
Pricing
Pay as you goBilled per run from your AI/ML API balance
- No subscription
- No seat fee
- One balance for every tool
Each request is billed per run from your AI/ML API balance, with no subscription or seat fee. Our test request for 20 reviews cost less than one cent, and so did the lookup that found the place. Paging through thousands of reviews takes many requests, so check the current price in your dashboard before a large pull.
FAQ
Is it legal to scrape Google reviews?
Reviews on Google Maps are public, and reading them for research or monitoring is common. Reviewer names and profile links are personal data under laws such as GDPR, so collect them only when you need them and store them with care. Google's terms apply to the service. Ask a lawyer about your specific case.
How many reviews can I get for one place?
Up to 100 on an unfiltered first request, then up to 20 per request with the page token from the previous response. Keyword and topic filters also return up to 20 at a time. Popular places have tens of thousands of reviews, so narrow the request with an order, a keyword or a topic instead of downloading everything.
Can I get only the negative reviews?
Sort by lowest rating and the one- and two-star reviews come first. Add a keyword such as “refund” to narrow them to one issue.
Do I need the Google place ID?
Not in a chat: name the place and the city and your assistant finds the ID with Google Maps Scraper first. In code, pass the place ID or Maps feature ID you stored from a Maps search.
Does it return the rating breakdown by stars?
Every response includes the place's average rating and total review count. The breakdown by stars, phone number and opening hours come from a Maps search of the place.
How is this different from the Google Places API?
The Places API returns at most five reviews per place and needs a Google Cloud project with billing. This tool returns as many reviews as you page through, sorted and filtered, and runs from AI assistants over MCP.
Can I reply to reviews with it?
No. It reads public reviews. Replies as the owner go through Google Business Profile for the locations you manage.
Can I use it from Claude, ChatGPT or Cursor?
Yes. Add the AI/ML API MCP server once and ask in plain English. The MCP quickstart covers the supported clients.
Can I export the reviews to CSV or Google Sheets?
Ask your assistant for the table as CSV, or write the JSON from the REST response to a file or a sheet in your code.
How fresh are the reviews?
Each run reads Google at that moment. In our test the newest review had been posted four hours before the run.
Read your first Google reviews
Connect once, name a place and ask what its customers say.