Booking.com Scraper

Get Booking.com hotels for a destination, or from search and hotel URLs: name, address, coordinates, price for your dates, rating with sub-scores, facilities, photos and anonymous guest reviews, one row per hotel. Revenue managers and travel startups use it to track prices and supply. $1 per 1,000 hotels.

Why this one

	Price per 1,000 results
This Actor	$1.00 per hotel ($0.50 per review)
Popular alternatives (top 5 on the Store by users)	$1.75 – $5.00

The top five Actors for "booking scraper" list between $1.75 and $5 per 1,000 results. Hotel rows and review rows are priced separately.

Quick start

Run the Actor with this input. Every other field has a default.

{
  "search": "Prague",
  "maxItems": 20,
  "currency": "EUR"
}

Output example

One row, trimmed. The full field list is further down this page.

{
  "recordType": "hotel",
  "name": "Hotel Residence Agnes",
  "stars": 4,
  "price": 278.71,
  "currency": "EUR",
  "nights": 2,
  "checkInDate": "2026-11-02",
  "rating": 9.9,
  "ratingLabel": "Exceptional",
  "reviews": 1277,
  "address": {
    "full": "Haštalská 943/19, Prague, 110 00, Czech Republic"
  }
}

Use it from your code

The Actor id is praguedatalab/booking-scraper. Replace YOUR_APIFY_TOKEN with your Apify API token.

Python (pip install apify-client)

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("praguedatalab/booking-scraper").call(run_input={"search":"Prague","maxItems":20,"currency":"EUR"})

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item)

JavaScript (npm install apify-client)

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });
const run = await client.actor('praguedatalab/booking-scraper').call({"search":"Prague","maxItems":20,"currency":"EUR"});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

curl (returns the rows as JSON; the endpoint waits for the run, up to 5 minutes, so keep the limits small)

curl -X POST "https://api.apify.com/v2/acts/praguedatalab~booking-scraper/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"search":"Prague","maxItems":20,"currency":"EUR"}'

Use it with AI agents

AI assistants can call praguedatalab/booking-scraper through the Apify MCP server. One call returns a list of hotels with prices for the dates you give.

FAQ

Is it legal to scrape Booking.com?

The Actor reads public hotel and review pages without logging in. It never collects reviewer names, pictures or countries. Booking.com's terms still apply to how you use the data. This is not legal advice.

How much do 10,000 results cost?

$10 for 10,000 hotels, or $5 for 10,000 reviews. A run that finds nothing charges nothing. Set a maximum cost per run to cap the spend.

Do I need a proxy?

Keep the default proxy setting: Apify Proxy with the RESIDENTIAL group. Booking.com blocks some IP addresses. If a run stops with a "blocked this IP" message, it charges nothing.

How fresh is the data?

Each run reads Booking.com when it runs. Prices are for the check-in and check-out dates in the input and change often. price is null when nothing is available for those dates.

Can I export to Excel, CSV or Google Sheets?

Yes. Each run saves a dataset that you can download as JSON, CSV or Excel. For Google Sheets, import the CSV, or connect Apify to Google Sheets with an integration.

Can I schedule it daily?

Yes. In the Apify Console, open the Actor, go to Schedules, and add a daily schedule with the same input. Each run is billed per result, as usual.

Pricing

Pay per result, nothing else (no platform usage added on top, no start fee):

Event	Price	You pay for
hotel	$0.001 ($1 per 1,000)	one hotel row
review	$0.0005 ($0.50 per 1,000)	one review row

Set a maximum cost per run in the Apify run options and the Actor stops cleanly when it is reached. A run that is blocked or finds nothing charges nothing.

Compatible with the input of voyager/booking-scraper

Switching takes one change: the Actor ID. The input names and values are the same: search, startUrls, maxItems, propertyType, sortBy, minScore, starsCountFilter, currency, language, checkIn, checkOut (a date or 14 days), rooms, adults, children, minMaxPrice. These two are accepted so existing inputs validate, but are not offered, and the run says so in its status message: extractAdditionalHotelData (surroundings, room offerings) and flexWindow.

The output keeps the leader's field names where they exist: order, url, startUrlOrQuery, name, type, description, stars, price, currency, rating, ratingLabel, reviews, breakfast, checkIn, checkOut, checkInDate, checkOutDate, location {lat, lng}, address {full, street, postalCode, country, region}, image, images, rooms, facilities, categoryReviews, hotelId, hotelChain, timeOfScrapeISO.

Differences to know:

Field	Here	Leader
currency	ISO code (EUR)	symbol as Booking.com shows it (€)
facilities	flat list of names	grouped by category
categoryReviews	[{name, score, count}]	similar, different keys
rooms	one entry per room offer with price, mealPlan, persons, bedType, features	similar
reviews	rows with recordType: "review" in the same dataset, no reviewer identity	separate Actor, with reviewer details
not provided	highlights, policies, roomImages, reviewSummary, licenseInfo, hostInfo, traderInfo, breadcrumbs, surroundings, roomOfferings, minimumNights	provided

Output

Hotel rows and review rows share one dataset (recordType tells them apart, and is the charged event). Every field of both is listed here and in the dataset schema; null means Booking.com did not show the value.

Field	Type	Row	Meaning
recordType	string	both	hotel or review. Also the charged event name.
order	integer	both	hotel: 1-based position in the search results that produced it (null for hotel URLs). review: position of the review among those saved for the hotel (newest first).
url	string	hotel	Booking.com hotel page with the stay dates, guests and currency of the run.
startUrlOrQuery	string	hotel	destination or start URL that led to this hotel.
hotelId	integer	both	Booking.com property id (hotel and review rows).
name	string	hotel	property name.
type	string	hotel	property type (hotel, apartment, aparthotel, hostel, guest_house, bed_and_breakfast, villa…).
description	string	hotel	property description from the hotel page (null when the hotel page was not opened).
stars	number	hotel	star classification 1-5, null when unrated.
price	number	hotel	lowest price for the whole stay (all nights, all rooms asked for) in currency; null when sold out for the dates.
pricePerNight	number	hotel	average price per night in currency.
currency	string	hotel	ISO 4217 code of price (the leader returns a symbol such as US$).
rating	number	both	hotel: guest rating on a 0-10 scale. review: this review's score (1-10).
ratingLabel	string	hotel	Booking.com's word for the rating (Exceptional, Superb, Fabulous, Very good, Good…); only when the hotel came from a search.
reviews	integer	hotel	number of guest reviews.
categoryReviews	array	hotel	review sub-scores (Staff, Facilities, Cleanliness, Comfort, Value for money, Location…) as {name, score, count}.
breakfast	string	hotel	breakfast styles offered, comma-separated.
checkIn	string	hotel	check-in time window, e.g. "From 15:00 to 22:00".
checkOut	string	hotel	check-out time window, e.g. "Until 11:00".
checkInDate	string	hotel	arrival date of the priced stay, YYYY-MM-DD.
checkOutDate	string	hotel	departure date of the priced stay, YYYY-MM-DD.
nights	integer	both	hotel: nights of the priced stay. review: nights the reviewer stayed.
location	object	hotel	coordinates as {lat, lng}.
address	object	hotel	address as {full, street, postalCode, city, region, country, countryCode}.
distanceFromCenter	string	hotel	distance from the city centre as Booking.com words it (search results only).
image	string	hotel	main photo URL.
images	array	hotel	gallery photo URLs (up to 100, 1024x768).
rooms	array	hotel	room offers for the stay as {roomId, roomType, bedType, persons, price, currency, mealPlan, available, features}; empty without availability.
facilities	array	hotel	every facility listed on the page, as a flat list of names.
popularFacilities	array	hotel	the "most popular facilities" shown on the hotel page.
hotelChain	string	hotel	brand or chain.
languagesSpoken	array	hotel	language codes spoken at the property.
freeCancellation	boolean	hotel	the cheapest offer in the search results can be cancelled for free (search results only).
isSponsored	boolean	hotel	the search result was marked as an ad.
finePrint	string	hotel	the fine print of the property.
timeOfScrapeISO	string	hotel	scrape time, ISO 8601 UTC.
hotelName	string	review	name of the reviewed property.
hotelUrl	string	review	hotel page URL.
reviewId	string	review	Booking.com's opaque review id.
date	string	review	date the review was written, YYYY-MM-DD.
title	string	review	review headline.
positiveText	string	review	what the guest liked.
negativeText	string	review	what the guest disliked.
language	string	review	language code of the text.
travelerType	string	review	couple, family, group of friends, solo traveller or business traveller.
stayStatus	string	review	stayed or similar status.
stayCheckIn	string	review	check-in date of the reviewed stay.
stayCheckOut	string	review	check-out date of the reviewed stay.
roomType	string	review	room the guest booked.
helpfulVotes	integer	review	helpful votes.
hotelReply	string	review	reply posted by the property.

Input

Field	Meaning
search	Destination, e.g. Prague.
startUrls	Booking.com search URLs (/searchresults…, their filters are kept) or hotel URLs (/hotel/cz/name.html). Dates, guests, currency and language of the input are applied on top.
maxItems	Hotels per destination or per search URL, at most 1,000 (higher values are capped). Default 20.
checkIn, checkOut	YYYY-MM-DD or relative (14 days, 2 weeks). Empty: next Friday, two nights. Prices and rooms need dates.
adults, children, rooms	Guests for the price.
currency, language	Currency (ISO code) and site language (en-gb, de, fr…).
propertyType, starsCountFilter, minScore, minMaxPrice, sortBy	Filters and order, as on Booking.com. minScore is applied exactly on the results (Booking.com itself only filters at 6, 7, 8 and 9).
extractDetails	Open every hotel page for description, facilities, photos, rooms and the full address (default on). Off gives a fast list from the search results only.
maxReviewsPerHotel	Anonymous reviews per hotel, newest first. 0 = none.
proxyConfiguration	Optional, see below.

How it works

Booking.com protects its pages with an AWS WAF JavaScript challenge. The Actor opens a real Chrome once, lets the challenge solve itself, keeps the aws-waf-token cookie, and then fetches everything else with a plain HTTP client that looks like Chrome: search pages, the next pages of a search (Booking.com's own GraphQL calls), hotel pages and reviews. When Booking.com asks for the challenge again, the Actor earns a new token. This is why the Actor runs with 2 GB of memory (the browser needs it for a few seconds) while the HTTP part is light.

Booking.com also answers some requests with a search page that holds no results at all (an empty "Redirect" answer, or a redirect to the search page without its query). The Actor never takes such a page for an empty destination: it asks again with the token cookie alone, then loads the search page in the browser itself, and only then gives up on that search. Every empty answer is logged in full (final URL, status, redirect chain, size, the verdict Booking.com wrote in the page, the first 500 characters) and the page is saved as SEARCH_PAGE_<n>.html in the run's key-value store, next to the SUMMARY record that lists all of them under searchDiagnostics.

Only public pages are used: no login, no captcha solving.

Proxy advice

Start without a proxy. Booking.com decides per IP address, so a run on a flagged address can be refused even though the same Actor works elsewhere. When that happens the run stops with the status message "Booking.com blocked this IP. Enable Apify Proxy with the RESIDENTIAL group…" and charges nothing. A search that comes back without results in every way the Actor can ask ends with the note "Booking.com returned no results for this search from this IP (…)", which names each attempt and what Booking.com answered; the same proxy advice applies. Enable it in the Proxy input: { "useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"] }. The browser and the HTTP client share one proxy session, so the token stays valid. Residential traffic is billed by Apify per GB; a hotel page is roughly 0.5 MB compressed, so turn off extractDetails when you only need the list.

Limits

• Booking.com shows at most 1,000 results per search. Split bigger destinations with filters (property type, stars, price ranges) or several searches.
• Reviews come 25 per request, newest first. The text is in the guest's language (language field); Booking.com's translation is not fetched.
• price is the lowest price for the whole stay and is null when nothing is available for the dates; rooms is empty then.
• Search ads (isSponsored: true) are kept and flagged. A hotel that several searches of one run find is saved once.
• Reviewer names, pictures and countries are never collected.
• The run keeps a progress record, so when Apify restarts a long run on another server it does not save or charge the same hotel twice (reviews of the hotel being processed at that moment can be incomplete).

Use cases

• Revenue managers and hotel groups: track competitor prices and ratings per night, per destination.
• Travel startups and aggregators: fill a destination database with coordinates, photos, facilities and live prices.
• Market research: map supply by property type and stars, compare rating distributions across cities.
• Reputation monitoring: pull the newest anonymous reviews of your own hotels and your competitors'.
• AI agents: one call returns a structured list of hotels with prices for given dates.

Terms and data

The Actor reads public Booking.com pages. Hotel names, addresses, prices and facilities are business data; review rows carry no reviewer identity. Check Booking.com's terms and the rules that apply to your use before running it at volume.