2GIS Search Scraper

Search public 2GIS results by the values a user would enter on the website and save a deduplicated, richly normalized business dataset.

What does this Actor do?

The Actor searches public 2GIS pages for business names, categories, services, and locations. It follows result pagination, merges duplicate businesses found by different searches or pages, and can enrich each result from the public Contacts, Info, Reviews, Photos, and Menu views.

It prefers structured JSON or server-rendered HTML when available and uses stable DOM text and links as a fallback. It never requires a result-page URL for normal searches, although public 2GIS search URLs are supported for advanced workflows.

Why use this Actor?

Use it when a plain list of names is not enough. Each record keeps source-backed contact, location, category, rating, hours, public web links, review, gallery, menu, and provenance fields in one stable contract. Multiple queries and start URLs can be combined into one dataset without duplicate businesses.

What data does it return?

Optional fields are omitted when 2GIS does not publicly expose them. The stable identity and provenance fields are always present on a valid record.

Field	Type	Description
id	string	Stable normalized 2GIS business identifier.
businessId	string	Business identifier when exposed separately by 2GIS.
name	string	Public business name.
address	string	Public listing address, assembled from the listing's own address block.
buildingName	string	Public building name shown with the business address.
streetAddress	string	Street and building label linked from the 2GIS address card.
floor	string	Public floor, level, suite, or unit label when shown.
addressArea	string	Neighborhood and administrative area line from the address card.
poBox	string	Public post office box number when shown.
city	string	Parsed city or region, when available.
categories	string[]	Public 2GIS categories.
matchedQueries	string[]	Searches that discovered the business.
matchedQueryIndexes	integer[]	Zero-based source indexes that discovered it.
matchedPages	object[]	Source page and query index pairs where it appeared.
rating	number	Public rating from 0 to 5.
ratingCount	integer	Public rating count.
reviewCount	integer	Public review count.
latitude	number	Public latitude.
longitude	number	Public longitude.
phoneNumbers	string[]	Public phone numbers.
workingHours	object	Day-to-hours mapping.
workingHoursText	string	Original public hours or opening-state text, including lunch-break notices.
openingStatus	string	Public open, closed, working, or time-to-change status.
websiteUrl	string	Sanitized public external website URL.
socialLinks	object[]	Public social platform and URL pairs.
branchesUrl	string	Public 2GIS branches URL.
branchCount	integer	Number of public branches when exposed.
routeUrl	string	Public 2GIS route URL.
description	string	Public page description from Open Graph or standard page metadata; it may be a 2GIS-generated summary.
averageBill	number	Public average bill value when exposed.
currency	string	Currency associated with the average bill.
seatingCapacity	integer	Public seating capacity when exposed.
cuisineTypes	string[]	Public cuisine types.
features	string[]	Public venue features.
attributes	string[]	Other public attributes.
assortment	string[]	Public assortment or product group labels.
services	string[]	Public services.
paymentMethods	string[]	Public payment methods.
serviceLanguages	string[]	Public service languages.
accessibility	string[]	Public accessibility information.
categoryLinks	object[]	Public related category names and URLs.
parkingOptions	object[]	Public parking type and space information.
landmarks	object[]	Public nearby landmark names, distances, and URLs.
relatedSearches	object[]	Public related search labels and URLs.
similarBusinesses	object[]	Public related business names, categories, listing URLs, and visible summary fields.
reviewSummary	object	Public review total and review-page URL.
reviewSnippets	object[]	Bounded public review text, rating, displayed reviewer name and review count, date, confirmation, visits, and useful count when visible. Profile identifiers and URLs are omitted.
photoCount	integer	Public gallery photo count.
photoCategories	object	Public gallery category counts.
photoUrls	string[]	Bounded stable public 2GIS photo URLs.
menuAvailable	boolean	Whether public menu photos are available; false when the Menu view shows its empty-menu prompt.
menuPhotoCount	integer	Public menu-photo count.
listingUrl	string	Canonical public 2GIS listing URL.
query	string	Source query associated with the output position.
queryIndex	integer	Zero-based source query or start-URL index.
page	integer	Result page associated with the output position.
sourceRank	integer	Stable first-discovery rank before final sorting.
position	integer	One-based final dataset position.
detailsFetched	boolean	Whether the public Contacts detail page completed.
source	string	Source label, currently 2GIS.
scrapedAt	string	ISO-8601 scrape timestamp.

How to scrape 2GIS

1. Enter a query, or enter several values in queries.
2. Optionally set city, filters, sorting, pagination, and enrichment limits.
3. Run the Actor and inspect the default dataset and the OUTPUT run summary.
4. Download or integrate the normalized records through the Apify API.

Minimal input

{
  "query": "restaurant",
  "city": "Dubai",
  "maxItems": 25,
  "maxPages": 2
}

Filtered and sorted input

{
  "query": "restaurant",
  "city": "Dubai",
  "sortBy": "ratingDesc",
  "minRating": 4,
  "minReviewCount": 10,
  "requireWebsite": true,
  "maxItems": 50,
  "maxPages": 3,
  "includeDetails": true,
  "includeReviews": true,
  "maxReviews": 3
}

Multiple searches and a starting page

{
  "queries": ["restaurant", "coffee"],
  "startUrls": [
    { "url": "https://2gis.ae/dubai/search/bakery/page/2" }
  ],
  "maxItems": 100,
  "maxPages": 2,
  "sortBy": "nameAsc"
}

The startUrls format is an array of objects with a required url, matching the Apify requestListSources editor. Only public HTTPS 2GIS search URLs on supported 2GIS domains are accepted.

Input

Search fields are query, queries, city, and startUrls. Limits are maxItems and maxPages. Ordering supports relevance, ratingDesc, ratingAsc, reviewCountDesc, reviewCountAsc, photoCountDesc, averageBillAsc, averageBillDesc, nameAsc, and nameDesc.

Available filters are includeKeywords, excludeKeywords, includeCategories, excludeCategories, minRating, maxRating, minReviewCount, minPhotoCount, minAverageBill, maxAverageBill, onlyOpenNow, requireWebsite, requirePhone, requirePhotos, requireMenu, and requireCoordinates.

Enrichment controls are includeDetails, includeInfo, includeReviews, includePhotos, includeMenu, maxReviews, and maxPhotos. Detail-dependent filters and sorts automatically enable the minimum required public detail pages.

Developer options are grouped in the input UI: maxConcurrency (1–8), maxRequestRetries (0–10), requestTimeoutSecs (5–120), requestDelayMs (0–10000), and datasetBatchSize (1–100). requestDelayMillis remains available as a compatibility alias for existing saved JSON inputs.

proxyConfiguration supports the Apify proxy editor, including authorized Apify Proxy or custom HTTP(S) proxy URLs. Direct requests are the default. Proxy credentials are never written to logs or records.

Chrome observation sample (not Actor output)

On 2026-09-23, the visible Chrome page 2GIS Dubai restaurant search was opened from the Actor's public URL https://2gis.com/search/restaurant and loaded without a challenge. It showed about 15,600 places and pagination through page 7. The following is a partial transcription of visible card fields. It documents the browser observation only; it is not an Actor run, dataset record, or extraction result.

[
  {
    "name": "HuQQabaz Garden",
    "businessId": "70000001055853809",
    "rating": 4.6,
    "ratingCount": 23,
    "address": "Mall of the Emirates / Sheikh Zayed Road / Al Barsha 1 / Dubai",
    "branchCount": 4,
    "averageBill": 200,
    "currency": "AED",
    "cuisineTypes": ["Turkish and Mediterranean cuisine"],
    "services": ["Delivery"]
  },
  {
    "name": "Reef Zam Zam",
    "categories": ["Kebab"],
    "branchCount": 12,
    "averageBill": 50,
    "currency": "AED",
    "cuisineTypes": ["Arabic cuisine"],
    "services": ["Delivery"],
    "workingHoursText": "Closes for lunch break in 13 minutes",
    "openingStatus": "Closes for lunch break in 13 minutes"
  }
]

Output

The Actor pushes only normalized records that match the dataset schema. Records are deduplicated by stable business ID, and matchedQueries, matchedQueryIndexes, and matchedPages preserve all known discovery provenance. Filters are evaluated after enrichment; non-relevance sorts are applied across the bounded combined candidate set before maxItems is enforced.

Example record:

For a complete record captured from a public 2GIS listing, see examples/sample-record.json.

{
  "id": "70000001092852304",
  "businessId": "70000001092852304",
  "name": "Example Cafe",
  "address": "1 Example Street",
  "city": "Dubai",
  "categories": ["Cafe", "Restaurant"],
  "matchedQueries": ["restaurant"],
  "matchedQueryIndexes": [0],
  "matchedPages": [{ "queryIndex": 0, "page": 1 }],
  "rating": 4.6,
  "ratingCount": 120,
  "reviewCount": 35,
  "latitude": 25.2048,
  "longitude": 55.2708,
  "phoneNumbers": ["+971 4 000 0000"],
  "workingHours": { "Monday": "09:00–23:00" },
  "openingStatus": "Open",
  "websiteUrl": "https://example.com",
  "cuisineTypes": ["International"],
  "features": ["Outdoor seating"],
  "services": ["Takeaway"],
  "paymentMethods": ["Cards"],
  "photoCount": 12,
  "photoUrls": ["https://i8.photo.2gis.com/images/branch/0/30258560077210013_7def.jpg"],
  "menuAvailable": true,
  "menuPhotoCount": 2,
  "listingUrl": "https://2gis.ae/dubai/firm/70000001092852304",
  "query": "restaurant",
  "queryIndex": 0,
  "page": 1,
  "sourceRank": 1,
  "position": 1,
  "detailsFetched": true,
  "source": "2GIS",
  "scrapedAt": "2026-09-05T12:00:00.000Z"
}

The Actor also writes a bounded OUTPUT key-value-store record containing status, item counts, pages, detail requests, filters, sorting, proxy use, dataset batches, and per-source stop reasons. You can download the dataset in various formats such as JSON, HTML, CSV, or Excel.

Pagination, filtering, and sorting

Each query and each start URL receives its own maxPages budget. A supplied /page/N URL continues at page N. Pagination stops on an empty result page, a reliable missing next-page link, the per-source page limit, or—when using default relevance without filters—the global maxItems limit. With filters or a non-relevance sort, the Actor continues through the bounded page set so the global filtered/sorted result is meaningful.

Search sources run concurrently up to maxConcurrency; pages within one source remain sequential. Detail views run in bounded concurrent workers. Dataset writes use datasetBatchSize and never exceed the global maxItems limit after deduplication and filtering.

How much does it cost?

Apify compute and proxy usage depend on the number of search pages, detail views, retries, and concurrency. Set includeDetails, maxPages, maxItems, maxReviews, and maxPhotos to control work. Direct requests are used unless an authorized proxy is selected.

Limitations and responsible use

Only publicly accessible 2GIS information is collected. The Actor does not log in, access private reviews, or bypass CAPTCHA or access controls. It retains only reviewer display names and review counts already shown publicly on the review card, and omits profile identifiers and URLs. Client-rendered gallery and menu content can be unavailable to the public HTTP response; those fields are omitted instead of guessed. A temporary block or changed website response is reported in OUTPUT and does not produce fabricated records.

Respect 2GIS terms, applicable law, robots and rate limits, and the rights of businesses and individuals. Use collected data for lawful purposes and avoid treating scraped information as verified business advice.

FAQ

Why is a rich field missing?

The field was not exposed in the public response for that business or view, or the value failed normalization. Optional fields are omitted so the dataset stays type-safe.

Why is my dataset empty?

Check OUTPUT for no-results, blocked, or failed. Filters intentionally reject records with missing values when a minimum or required field was requested. Start with a simple query and no filters, then add constraints gradually.

Can I reuse a saved input JSON file?

Yes. The repository includes working examples in qa-inputs/, and requestDelayMillis remains supported for older saved inputs. The Actor validates input bounds before network work begins.

Where do I see failures?

Open the OUTPUT key-value-store record. It contains bounded counts and per-source stop reasons without storing authorization headers, cookies, proxy credentials, signed URLs, or raw response bodies.

Support

For a reproducible issue, include the sanitized input shape, Actor run ID, OUTPUT summary, and the affected public 2GIS region/query. Do not include API tokens, cookies, proxy credentials, or private account data. This Actor is an independent tool and is not affiliated with 2GIS.