2GIS Reviews Scraper

2GIS Reviews Scraper: every review on a place, with the owner's reply

Paste 2GIS place pages, or search a city, and get one row per review: the star rating, the full text as the author wrote it, who wrote it, when, any photos they attached and the business owner's reply. Every row also carries the place it belongs to, down to the address and coordinates.

The one thing to know before you set a big number: 2GIS serves every review a place has ever collected, so a busy chain branch can run to thousands. Set the per-place cap first and the total cap second.

💬 What 2GIS Reviews Scraper does

Two ways to name the places, and they work together in one run.

URLs and ids. Paste a place page, the reviews tab of a place, or just the bare numeric place id. A 2GIS search page works too, and every place that search finds gets scraped.

Search terms. Give it a term and a city instead, and it finds the places first, then reads the reviews on each one. maxPlaces is what stops that from running away.

Reviews come back newest first by default. 2GIS has no rating sort of its own, so if you only want the angry ones, pull them in date order and filter on rating afterwards, or use Minimum rating to keep only the stars you care about.

📥 What you give it

{
  "startUrls": ["https://2gis.ru/moscow/firm/70000001044357973"],
  "maxReviewsPerPlace": 200,
  "maxItems": 500,
  "onlyWithText": true
}

Run it with everything empty and you get one labelled sample row.

📤 What you get back

A real review row from a recent run. The owner's reply ended with an emoji, cut short here:

{
  "reviewId": "294305879",
  "rating": 5,
  "text": "Отличная пиццерия, посидела с парнем все понравилось, было очень вкусно, приятная атмосфера, самая вкусная пицца в городе, вежливый персонал, буду приходить еще",
  "textLength": 160,
  "authorName": "Алина Сидоренко",
  "authorId": "108196462",
  "authorPublicId": "d8b4a52856c24a4baf71635c87d9b68c",
  "authorReviewCount": 1,
  "authorIsVerified": false,
  "dateCreated": "2026-09-05T09:58:14.612353+07:00",
  "dateEdited": null,
  "likesCount": 0,
  "commentsCount": 1,
  "photoCount": 0,
  "photoUrls": [],
  "officialAnswerText": "Спасибо! Круто знать, что вы довольны. Постараемся и дальше только радовать ...",
  "officialAnswerDate": "2026-09-05T11:20:32.795509+07:00",
  "officialAnswerBy": "Додо Пицца, сеть пиццерий",
  "source": "2gis",
  "placeId": "70000001044357973",
  "placeName": "Додо Пицца, сеть пиццерий",
  "placeCategory": "Пиццерии",
  "placeAddress": "Москва, Театральный проезд, 5",
  "placeCity": "Москва",
  "placeCountry": "Россия",
  "placeLatitude": 55.759891,
  "placeLongitude": 37.625011,
  "placeRating": 4.9,
  "placeReviewCount": 296,
  "placeUrl": "https://2gis.ru/moscow/firm/70000001044357973",
  "reviewUrl": "https://2gis.ru/moscow/firm/70000001044357973/tab/reviews",
  "position": 1,
  "scrapedAt": "2026-09-14T05:41:47.687Z"
}

🧾 Reading the output

Three kinds of row can land in your dataset.

Diagnostic rows are never charged, and neither are reviews your filters dropped.

▶️ How to run it

1. Open 2GIS Reviews Scraper and click Try for free. 2. Paste your place links into 2GIS place URLs, one per line. 3. Set Maximum reviews per place, then Maximum reviews in total as your spend cap. 4. Click Start. 5. Download the dataset as JSON, CSV or Excel, or read it from the Apify API.

💰 How much does it cost?

$0.50 per 1,000 reviews. Flat on every Apify plan, no volume tiers.

One row is one review. The sample row, every diagnostic row and every review your filters dropped are not charged. A place with no reviews produces no rows to charge for.

💡 What people use it for

• Reading what customers say about a competing branch before opening nearby.
• Tracking a chain's rating across every branch in a city, week by week.
• Pulling one-star and two-star reviews to find the complaint that keeps repeating.
• Checking which branches answer their reviews and which ignore them, through officialAnswerText.
• Feeding review text into a summariser, with the place details already on each row.

🚧 What it does not do

• No rating sort. 2GIS orders by date or by its own trust score, nothing else. Minimum rating

filters after the fact, it does not reorder.

• The per-place cap counts what it reads, not what it keeps. With Minimum rating or

Only reviews with text switched on, a place can return fewer rows than you asked for even though more matching reviews exist further down. Raise maxReviewsPerPlace when you filter hard.

• Around 5,000 reviews in one run. The box will accept 100,000, but rows pile up in memory and a

long run slows down badly. Split a big job across several runs.

• Paste more places than maxPlaces and the extras are dropped quietly, with no row to tell you.

Raise the cap to match your list.

• A bad city stops everything, including any place URLs you pasted in the same run, whenever

search terms are in the input too.

• A hiccup on the first page loses that place. If 2GIS stumbles partway, you get a diagnostic for

the place rather than a partial list. Re-run just that one.

• Photo links, not photo files. photoUrls points at 2GIS; nothing is downloaded.
• Display names only. No reviewer emails, no phone numbers. 2GIS does not publish them.

🧭 Which reviews scraper do you need?

❓ Questions people ask

Do I need a 2GIS account or an API key? No. Nothing to register and no quota to watch.

Which countries have reviews? The ones 2GIS maps: Russia, Kazakhstan, Uzbekistan, Kyrgyzstan, Azerbaijan, Belarus, Georgia, Armenia, Tajikistan, Cyprus and the UAE.

Can I get only the bad reviews? Not directly from 2GIS. Pull them in date order and filter on rating in your own tool, which also lets you keep the text you already paid for.

Is the review text translated? Never. It comes back in whatever language the author used, and locale only affects place names and categories.

Can I run it on a schedule? Yes. reviewId is stable, so a scheduled run plus a dedupe on that field gives you only what is new.

Is scraping 2GIS reviews legal? These are public reviews on public business pages. Apify's write-up on scraping and the law is a good place to start, and we are not lawyers.

🆘 If something breaks

Open the Issues tab on the actor page. Send the place URL and the run ID. The errorCode on the diagnostic row usually names the problem on its own.