Google Maps Reviews Scraper & Sentiment Analyzer

Scrape and analyze public Google Maps reviews with multilingual sentiment classification. Submit up to 10 business URLs and receive one structured Dataset item per business with positive, neutral, and negative distributions plus review-level confidence values.

What this Actor does

Analyze public Google Maps reviews from one or more business listing URLs and receive one compact sentiment report per business. The Actor resolves each business over HTTPS, collects a bounded set of unique reviews, classifies each review as positive, neutral, or negative, and returns counts, percentages, confidence values, and per-review labels.

The default analysis uses the open-source clapAI/mmBERT-small-multilingual-sentiment checkpoint. It is designed for multilingual sentiment classification and supports the languages documented by its model card. The Actor does not use a generative LLM, does not create business-specific categories, and does not write narrative summaries.

The model runs as reviews are collected. If a review has no text, its rating is used as a clearly marked fallback. The output identifies whether each result came from the model or that fallback.

Use cases

• Measure the positive, neutral, and negative share of a local business's reviews.
• Compare text sentiment with the star ratings shown on Google Maps.
• Feed review sentiment into SEO reports, CRMs, dashboards, or monitoring workflows.
• Process businesses in different countries without maintaining a category list for each industry.
• Retain only review identifiers, ratings, sentiment, and confidence instead of reviewer profiles or raw text.

How to use

1. Open the Actor in Apify Console.
2. Add one business URL on one line, or add several lines for a batch.
3. Set the maximum number of reviews per business.
4. Click Start and open the default Dataset when the run finishes.

Input

The Actor has one URL field: urls. It supports both ways users normally run the Actor without making them choose between overlapping inputs:

• One business: add one Google Maps business URL on one line. The Actor returns one Dataset item.
• Several businesses: add one URL per line, up to 10 different businesses. The Actor processes them in the same order and returns one Dataset item per business.

Do not add a separate url property. The only accepted URL property is the urls array, even when it contains just one entry.

Every entry must be a complete HTTPS Google Maps business listing URL. The following are the two most common supported formats. They are alternatives: choose the format you have and always place it inside the same urls array.

Example 1: Google Maps URL with a Place ID

Use this format when you already have a Google Place ID:

{
  "urls": [
    "https://www.google.com/maps/place/?q=place_id:ChIJ--511_65pBIR4Tl2-D4grUw"
  ],
  "maxReviews": 5,
  "sort": "recent",
  "language": "es",
  "country": "ES",
  "includeRawReviews": false
}

Example 2: Full Google Maps business listing URL

Use the complete URL copied from an open Google Maps business listing. Long parameters such as coordinates and /data=... are supported:

{
  "urls": [
    "https://www.google.com/maps/place/Walmart+Supercenter/@40.6607799,-73.726629,17z/data=!3m1!4b1!4m6!3m5!1s0x89c264150e65d803:0x17efe5921b20790f!8m2!3d40.6607799!4d-73.726629!16s%2Fg%2F1tdkhkzl?authuser=0&hl=en&entry=ttu&g_ep=EgoyMDI2MDgxOS4wIKXMDSoASAFQAw%3D%3D"
  ],
  "maxReviews": 5,
  "sort": "recent",
  "language": "en",
  "country": "US",
  "includeRawReviews": false
}

Standard /maps/place/ URLs, maps.app.goo.gl shared links, goo.gl/maps links, and supported Maps URLs containing place_id, cid, or ludocid identifiers are also accepted. Normal Google Search URLs, Maps searches that do not identify one business, directions, user contribution/profile URLs, non-Google domains, HTTP links, empty values, and duplicate URLs are rejected before collection starts.

For a batch, add up to 10 different business URLs to the same urls array. You may mix supported URL formats in one run. Do not paste the two complete JSON examples one after another; submit one JSON object containing all the URLs you want to analyze.

maxReviews defaults to 100 and accepts 1 to 500 unique reviews per business. sort defaults to recent and accepts relevance, recent, highest, or lowest. language and country localize the Google Maps request and default to es and ES; they do not restrict the language of the review text.

Sentiment is always analyzed with the bundled multilingual model. The Actor processes review text as it is collected; rows without text use the star rating as an explicitly marked fallback. includeRawReviews defaults to false; when enabled, the result also includes normalized review IDs, dates, timestamps, ratings, and text. Author names, profiles, avatars, and profile URLs are not included.

The complete output example below was produced from the full business listing URL shown in Example 2.

Output

The default Dataset contains one aggregate item per unique submitted business URL. place.inputUrl identifies the business that produced the row. place contains the resolved Place ID, business name, rating, total review count, and one-to-five-star distribution when Google exposes them. collection reports the customer-facing collection status and review coverage, sentiment contains the aggregate values, and reviewSentiments contains one compact classification per collected review without raw text. Internal transport diagnostics are not included in Dataset items. Businesses are processed in input order and each completed row is pushed independently.

The sentiment summary contains distribution, percentages, averageConfidence, reviewsAnalyzed, textReviewsAnalyzed, reviewsWithoutText, modelUsed, fallbackUsed, and warnings. Each item in reviewSentiments contains reviewId, rating, sentiment, confidence, source, and textAvailable.

This is a complete Dataset item from a successful analysis:

{
  "place": {
    "inputUrl": "https://www.google.com/maps/place/Walmart+Supercenter/@40.6607799,-73.726629,17z/data=!3m1!4b1!4m6!3m5!1s0x89c264150e65d803:0x17efe5921b20790f!8m2!3d40.6607799!4d-73.726629!16s%2Fg%2F1tdkhkzl?authuser=0&hl=en&entry=ttu&g_ep=EgoyMDI2MDgxOS4wIKXMDSoASAFQAw%3D%3D",
    "resolvedUrl": "https://www.google.com/maps/place/Walmart+Supercenter/@40.6607799,-73.726629,17z/data=!3m1!4b1!4m6!3m5!1s0x89c264150e65d803:0x17efe5921b20790f!8m2!3d40.6607799!4d-73.726629!16s%2Fg%2F1tdkhkzl?authuser=0&hl=en&entry=ttu&g_ep=EgoyMDI2MDgxOS4wIKXMDSoASAFQAw%3D%3D",
    "placeId": "ChIJA9hlDhVkwokRD3kgG5Ll7xc",
    "name": "Walmart Supercenter",
    "rating": 3.9,
    "totalReviews": 24133,
    "ratingDistribution": {
      "1": 2575,
      "2": 1378,
      "3": 2952,
      "4": 5162,
      "5": 12066
    }
  },
  "collection": {
    "status": "ok",
    "requested": 5,
    "collected": 5,
    "uniqueCollected": 5,
    "sort": "recent"
  },
  "sentiment": {
    "mode": "multilingual_model",
    "modelUsed": true,
    "model": "clapAI/mmBERT-small-multilingual-sentiment",
    "reviewsAnalyzed": 5,
    "textReviewsAnalyzed": 5,
    "reviewsWithoutText": 0,
    "fallbackUsed": false,
    "distribution": {
      "positive": 3,
      "neutral": 2,
      "negative": 0
    },
    "percentages": {
      "positive": 60,
      "neutral": 40,
      "negative": 0
    },
    "averageConfidence": 0.778575,
    "warnings": []
  },
  "reviewSentiments": [
    {
      "reviewId": "Ci9DQUlRQUNvZENodHljRjlvT25oUE9YWkdlVkJzVm13eVoyNTVYM014WlhScGMxRRAB",
      "rating": 5,
      "sentiment": "neutral",
      "confidence": 0.473044,
      "source": "multilingual_model",
      "textAvailable": true
    },
    {
      "reviewId": "Ci9DQUlRQUNvZENodHljRjlvT2tnd1oyMUdiMmQxU3pBek1rYzNjMlZXVFhkdVNGRRAB",
      "rating": 4,
      "sentiment": "neutral",
      "confidence": 0.695602,
      "source": "multilingual_model",
      "textAvailable": true
    },
    {
      "reviewId": "Ci9DQUlRQUNvZENodHljRjlvT21ab1owRnRhWFoxZDFCb2VuSjFiVkpzUzBoQ2JVRRAB",
      "rating": 5,
      "sentiment": "positive",
      "confidence": 0.794851,
      "source": "multilingual_model",
      "textAvailable": true
    },
    {
      "reviewId": "Ci9DQUlRQUNvZENodHljRjlvT2t0aGRHSjVSRE15UVVkd1QxRlJUalJJZUhkRFNXYxAB",
      "rating": 5,
      "sentiment": "positive",
      "confidence": 0.959329,
      "source": "multilingual_model",
      "textAvailable": true
    },
    {
      "reviewId": "Ci9DQUlRQUNvZENodHljRjlvT25wUVJVTmhURTl2UkRnNE1IcGxRWFYzZWtoZlExRRAB",
      "rating": 5,
      "sentiment": "positive",
      "confidence": 0.970048,
      "source": "multilingual_model",
      "textAvailable": true
    }
  ],
  "quality": {
    "textCoverage": 1,
    "ratingCoverage": 1,
    "absoluteTimestampCoverage": 0,
    "deduplicated": true
  },
  "analysis": {
    "mode": "multilingual_model",
    "llmUsed": false,
    "method": "basic_multilingual_sentiment",
    "reviewsObserved": 5,
    "warnings": [],
    "modelUsed": true,
    "limitations": [
      "Sentiment is limited to positive, neutral, and negative labels.",
      "Rows without text use the star rating as a fallback.",
      "The model confidence is not a calibrated probability."
    ]
  }
}

Results can be inspected in the Dataset or exported in the formats supported by Apify. Review text is not returned unless includeRawReviews is enabled.

Pricing

This Actor uses pay-per-event pricing. Apify automatically applies the rate for your account tier:

Apify tier	Price per analyzed review	Price per 1,000 reviews
FREE	$0.001	$1.00
BRONZE	$0.0009	$0.90
SILVER	$0.0008	$0.80
GOLD	$0.00075	$0.75
PLATINUM	$0.00075	$0.75
DIAMOND	$0.00075	$0.75

There is no additional charge per business, Actor start, or Dataset item, and platform usage is included in the event price.

A review-analyzed event is charged only for a review included in a delivered Dataset result. Zero-review, blocked, unresolved, and error results are not charged. If a business yields fewer reviews than requested, only the reviews actually analyzed and delivered are charged.

You can set a maximum total charge for the run in Apify. The Actor checks the remaining paid capacity before collecting each business and may return fewer reviews when that limit is reached, so the run does not intentionally analyze more paid reviews than the configured budget allows.

API and integrations

Run the Actor from Apify Console, the REST API, Apify clients, CLI, Tasks, Schedules, or webhooks. Use place.placeId when it is available as a stable key for comparing repeated runs. Use sentiment.distribution and sentiment.percentages for aggregate reporting, or reviewSentiments for row-level integrations.

Related Actors

Actor	Best for
Google Maps Business Scraper | $1 per 1,000	Discovering Google Maps business listings by location and category before selecting individual businesses for review analysis.
Google SERP Scraper API – Search Results Scraper	Comparing review sentiment with the business's organic Google visibility for selected searches.
Similarweb Website Traffic & SEO Analytics Scraper	Adding estimated traffic and SEO context to a business website found from its Google Maps listing.

Limits and data quality

• One run requires urls with 1 to 10 different Google Maps business URLs and accepts up to 500 requested reviews per business.
• The Actor processes businesses sequentially and reuses one sentiment-model instance; it does not multiply model memory by the number of URLs.
• Google Maps URLs and response structures can change. The result reports statuses such as partial, blocked, not_found, or ambiguous when collection cannot be verified.
• collected is the number of unique reviews returned by the source, not a guarantee that it equals the business's total review count.
• The bundled multilingual model classifies available review text. Rows without text are counted with their rating and marked with source: "rating_fallback".
• The model returns three sentiment classes. confidence is the model's top-class score and is not a calibrated probability or a guarantee of correctness.
• A review can contain mixed opinions, sarcasm, multiple languages, or culturally specific wording. Validate results before using them for high-impact decisions.
• The rating fallback is deterministic and uses only star ratings for reviews where no text is available; it does not replace the model analysis.
• Raw reviews are disabled by default. When enabled, review text may contain personal information written by third parties.
• Source availability, localization, rate limits, and network conditions can affect review coverage.

Frequently asked questions

Do I need to provide a Google place_id?

No. Provide the normal public business listing URL. The Actor attempts to resolve the available business identifier over HTTPS. A URL that does not identify one unique listing can return an ambiguous or not_found status.

Does this Actor use a browser or generative LLM?

No. Review extraction uses direct HTTPS requests and the single analysis flow uses a local open-source text-classification model. The output explicitly reports llmUsed: false.

How is sentiment analyzed?

The Actor uses one bundled multilingual text-classification model and returns positive, neutral, or negative with a confidence score. Reviews without text use their star rating as a clearly marked fallback.

Can I analyze multiple businesses in one run?

Yes. Add one URL per line in urls, up to 10 different businesses. The Dataset receives one independent result per URL, in input order. If one business cannot be resolved or collected, its status is reported in that row and the remaining URLs continue processing.

Why did I receive fewer reviews than requested?

Google Maps may expose fewer reviews in the selected order, stop returning additional reviews, or temporarily restrict a request. The Actor can also cap collection at the run's maximum total charge. Inspect collection.status, collection.requested, and collection.collected instead of treating a short result as proof that the business has no more reviews.

How do I limit the maximum cost?

Set the maximum total charge when starting the run in Apify. The number of reviews covered depends on your account tier; for example, a $0.10 limit covers up to 100 reviews on FREE or more on a discounted tier. When the remaining budget cannot cover another review at the safest applicable rate, the Actor stops before performing additional paid collection.

Are individual reviews included by default?

The default output includes compact review IDs, ratings, sentiment, and confidence. Raw review text and timestamps are included only when includeRawReviews is enabled.

Responsible use

Use this Actor only for lawful purposes and in accordance with applicable laws, privacy obligations, database rights, and the terms governing Google Maps data. You are responsible for determining whether collecting, storing, analyzing, and redistributing review content is permitted for your use case.

Minimize retained personal data, protect exported Datasets, avoid excessive collection, and honor applicable deletion or opt-out requests. Review text can contain personal information written by third parties even when author fields are omitted.

Support

If a run returns an unexpected status or output, open an issue in the Actor's Issues tab with the run ID, a sanitized input URL, collection.status, and the observed behavior. Do not post API tokens, cookies, credentials, or private review content.