Booking.com Scraper

Scrape Booking.com accommodation listings from public city pages or search-results URLs and export Booking.com property data to JSON, CSV, or spreadsheet formats. This Booking.com API alternative collects property names, page links, review scores, locations, images, and any public price and room information exposed by the selected page.

A city URL is a dependable starting point when you need a compact catalogue of properties. A dated destination search reads the public search data response and can provide prices and room summaries. Search-results URLs with website filters use the data embedded in their pages and may require visitor verification. Each dataset row represents one unique property, identified by its public property ID.

What data can you extract from Booking.com?

The actor keeps the same columns for search and city records. Empty values mean the source page did not expose the field. A city page usually includes names, review scores, images, and location; a dated search page can include prices, room summaries, and more amenities. The sourcePage column tells you which page supplied a row.

Field	Description	Example
order	Position in this run's output	1
propertyId	Public property identifier used for deduplication	7886254
url	Direct public property page	https://www.booking.com/hotel/nl/tribe-amsterdam-city.html
name	Displayed accommodation name	Tribe Amsterdam City
type	Property type label, if exposed	null
propertyTypeId	Numeric accommodation category, when present	204
description	Short public property summary	Located in Amsterdam...
stars	Star count where the page marks the score as stars	4
price	Numeric displayed amount in currency	103.84
pricePerNight	Average nightly amount on a dated search card	82.00
priceBasis	from on a city card or totalStay on a dated search card	from
currency	Three-letter code attached to price	USD
rating	Guest score on the site's ten-point scale	9
ratingLabel	Displayed score label, when present	Very Good
reviews	Number of guest reviews shown on the card	5811
breakfast	Meal-plan inclusion indicated on a search card	null
checkIn	Property check-in time if the card exposes it	null
checkOut	Property check-out time if the card exposes it	null
checkInDate	Date applied to a search result	2026-11-10
checkOutDate	Departure date applied to a search result	2026-11-12
latitude	Property latitude where exposed	52.402493
longitude	Property longitude where exposed	4.931046
address	Street address on a search card, when available	2 Middenweg
city	City name	Amsterdam
countryCode	Two-letter country code	nl
district	District or displayed area	Amsterdam Noord
image	Main public property image URL	https://cf.bstatic.com/xdata/images/hotel/...
features	Amenity names shown on a search result	["Free Wifi"]
rooms	Room summaries available on a dated search result	[{"roomType":"Original Room","persons":2}]
isSoldOut	Availability marker on a search result	false
sourcePage	Source of the record	city

The rooms array holds the room name, occupancy, displayed amount and currency, bed count, and bedroom count when the search card provides them. It is a summary of visible room information, not a full rate plan or reservation quote. The public city collections do not include room arrays, so rooms is empty for those rows. features is similarly an empty array when no amenity names appear. Prices and availability can change between page retrieval and a booking decision.

How to scrape Booking.com

1. Copy a public city URL, such as https://www.booking.com/city/nl/amsterdam.html, into startUrls. You may provide several city URLs in one run. Alternatively, enter a destination in search or paste a search-results URL containing the filters you selected on the website.
2. Set maxItems to the number of unique properties you need. Add checkIn and checkOut when using search results and you want the page to expose stay pricing. Choose the number of adults and rooms for that search.
3. Select a preferred currency and language if needed. Existing URL query parameters take precedence over these inputs, so a filtered search URL retains the choices already present in it.
4. Run the actor and inspect the dataset. Download the rows in JSON, CSV, or a spreadsheet format. Check sourcePage and priceBasis before comparing prices from different inputs.

For a compact list of names and ratings, start with city pages. For comparison of prices for a particular stay, use a dated destination search. A dated search with no website-specific filters uses the public search data response; a URL containing additional filters uses its embedded page data. The actor does not open each hotel's detail page. A hotel-detail URL by itself is not an accepted input.

Input

Name	Type	Default	Description
search	text	empty	Destination name used for a search-results request.
startUrls	URL list	empty	Public city or search-results URLs. Search URL filters remain in the request.
maxItems	integer	100	Maximum unique properties across all input URLs; 0 removes the actor-side cap.
checkIn	date	empty	Arrival date for a search, formatted YYYY-MM-DD.
checkOut	date	empty	Departure date after checkIn; both dates must be supplied together.
adults	integer	2	Adult guests for a search-results request.
rooms	integer	1	Rooms for a search-results request.
currency	text	USD	Preferred three-letter displayed price currency.
language	text	en-us	Preferred page language code.
proxyConfiguration	object	automatic routing	Optional proxy configuration.

Example for a dependable catalogue run:

{
  "startUrls": [
    { "url": "https://www.booking.com/city/nl/amsterdam.html" },
    { "url": "https://www.booking.com/city/fr/paris.html" }
  ],
  "maxItems": 50,
  "currency": "USD",
  "language": "en-us"
}

For dated results, use a search URL or a destination name along with two dates. For example, search can be Amsterdam, checkIn can be 2026-11-10, and checkOut can be 2026-11-12. Those date strings are examples; choose future dates appropriate to your run. A search URL can also carry price, star, property-category, and sorting choices made on the website. The actor preserves existing query parameters rather than silently replacing them.

You can mix city and search URLs, but rows are deduplicated by public property ID. Put the source with the fields you care most about first. If a property appears in both sources, the first version is retained. For example, place a dated search-results URL before a city URL if the dated amount matters more than the short city description.

Output example

The following item was collected from the public Amsterdam city page in the cloud validation run. The image URL is shortened here for readability; the dataset contains the full URL. The value priceBasis: "from" matters: this city-card amount is an indicative displayed price, not a quote for the example dates above.

{
  "order": 1,
  "propertyId": 7886254,
  "url": "https://www.booking.com/hotel/nl/tribe-amsterdam-city.html",
  "name": "Tribe Amsterdam City",
  "type": null,
  "propertyTypeId": 204,
  "description": "Located in Amsterdam, 2.2 miles from A'DAM Lookout, Tribe Amsterdam City has accommodations with a fitness center, private parking, a shared lounge and a bar.",
  "stars": null,
  "price": 103.84,
  "pricePerNight": null,
  "priceBasis": "from",
  "currency": "USD",
  "rating": 9,
  "ratingLabel": null,
  "reviews": 5811,
  "breakfast": null,
  "checkIn": null,
  "checkOut": null,
  "checkInDate": null,
  "checkOutDate": null,
  "latitude": 52.402493,
  "longitude": 4.931046,
  "address": null,
  "city": "Amsterdam",
  "countryCode": "nl",
  "district": "Amsterdam Noord",
  "image": "https://cf.bstatic.com/xdata/images/hotel/...",
  "features": [],
  "rooms": [],
  "isSoldOut": null,
  "sourcePage": "city"
}

In that run, two city URLs yielded 50 unique rows before the configured item limit stopped collection. Names, public property IDs, links, review scores, review counts, city names, and images were present on all 50 rows. Some source cards omitted the country code; the actor now derives it from the public hotel URL when needed. A separate dated Amsterdam search validation delivered 30 rows with dates, names, links, ratings, and review counts present on every row. The example preserves empty fields rather than filling them with estimates.

Use cases

• Accommodation market research: collect public property names, categories, review counts, and locations for a city-level inventory.
• Destination analysis: compare public review scores and listing counts across several city pages using a single dataset.
• Trip planning: shortlist properties from a filtered search URL, then open the linked property pages to confirm the current offer.
• Location analysis: map listings with available coordinates or group them by city, district, and country code.
• Rate monitoring: run the same dated search at intervals and compare displayed amounts, keeping the dates, currency, and priceBasis with each snapshot.
• Content audits: identify listings with missing public descriptions, ratings, images, or other fields for follow-up checks.

A record reflects what the public page showed at retrieval time. It is suitable for research and monitoring; it does not reserve inventory, hold a rate, or confirm that every room remains bookable. When you combine runs, keep the input dates and page URL alongside the exported data so price comparisons remain interpretable.

How much does it cost to scrape Booking.com?

The current result charge is $1.00 per 1,000 dataset items. At that rate, 50 delivered properties cost about $0.05, 500 properties about $0.50, and 5,000 properties across suitable inputs about $5.00. These are calculations from the per-result rate, not promises that one city page or one search can supply those volumes. The platform also enforces the run's own spending limit, and the actor stops pushing rows when the item cap or charge limit is reached.

A city page exposes a selected group of properties, so an input with maxItems: 500 may return far fewer than 500 items. The Amsterdam and Paris city pages in the validation run supplied enough distinct properties to reach the configured 50-item limit together. Search pages can show more results but may be blocked by visitor verification. Inspect the actual dataset count after each run; charges depend on items delivered, not the requested maximum.

Tips for best results

Use city URLs when you need names, links, ratings, photos, and a modest sample of properties. They are useful for a stable first pass and can be combined across destinations. City collections may include the same hotel in several page sections; the actor removes duplicates by propertyId before writing the dataset.

Use a search-results URL if you need website filters or a stay-specific amount. Apply property category, star, price, or sorting choices on the website, then copy the resulting URL. The actor keeps those query parameters. Add dates to see more complete pricing and room information where the page provides it. Without stay dates, a result may have no meaningful price or room match.

If a price is essential, require a nonempty price, currency, and appropriate priceBasis in your downstream workflow. A city-card from amount and a dated-search totalStay amount have different meanings. Do not compare them as equal nightly prices. pricePerNight is available only when the dated search card includes it.

Set a realistic item cap. Public search pagination commonly exposes only part of a destination's inventory, and the actor stops advancing after 1,000 search offsets per URL. To cover more inventory, submit several separately filtered search URLs and allow deduplication across them. Splitting by property category, district, or price range may expose different results, subject to what the website displays.

Limits

The search-results page and hotel-detail routes sometimes respond with a JavaScript verification page. A browser attempt and a rotated residential session did not consistently clear that response during development. Dated destination searches use the public search data response and passed the 30-item validation. Filtered search URLs still depend on their page data and can fail verification despite retries. City landing pages returned normal property data in the validation run. A run with no collected properties fails visibly instead of returning an apparently successful empty dataset.

City pages are curated sections, not an exhaustive city inventory. In the observed Amsterdam page, its public sections exposed 34 distinct property IDs; the observed Paris page exposed 33. Those counts may change. Search results also have a visible per-query limit, and a single query should not be treated as a complete catalogue. maxItems: 0 removes only this actor's output cap; it cannot remove site limits.

Hotel check-in and check-out times, complete addresses, detailed room offers, and full descriptions are usually absent from the city collections. The output retains their columns as null or empty arrays. Search cards provide some additional fields but are not hotel-detail pages. Direct hotel URLs are not accepted as startUrls. No login, personal account data, or private guest details are collected.

Language, currency, date, and location can affect what the site shows. If a supplied URL already sets one of those query parameters, the URL value is preserved. A city's indicative amount may not be tied to the input dates; dates apply to search pages. For any purchase decision, reopen the hotel link and verify the current price, taxes, conditions, and availability on the website.

Use with connected clients

A connected client can run the actor with a city URL and request a table of property name, district, guest score, review count, and link. Another useful request is to run a dated search URL for two guests, return properties with a displayed total-stay amount, and sort the exported rows by review score. Include the requested dates and currency in the client's task so the result can be interpreted correctly.

The input descriptions specify accepted URL types, limits, dates, and currency. The output includes a source marker and price basis, so a downstream application can separate city catalogue rows from dated search rows. A client should handle absent fields as missing source data rather than treating null as zero.

FAQ

Is it legal to collect these listings? The actor reads public accommodation pages. Use the data in a way that respects applicable laws, the website's terms, and privacy rules. Do not use the output to infer private guest information.

Do I need to configure a proxy? Usually no. The default setup starts with datacenter routing and can change session or escalate when a request receives a verification response. A custom proxy setting is available for accounts with different network requirements.

How fast is a run? The validated 50-row city run finished in seconds after startup. Search runs can take longer because each page needs a request and verification retries may occur. Performance also changes with the number of URLs and result pages.

Can I schedule repeat runs? Yes. Reuse the same input at a chosen interval and store each dataset with its run time. For price monitoring, keep the exact dates and currency in the input, and compare the same priceBasis across runs.

Can I export to a spreadsheet? Yes. Download the dataset as CSV or a spreadsheet-compatible format. Nested room summaries remain structured in JSON; for a simple table, use top-level columns such as name, rating, price, currency, city, and URL.

Why are some columns empty? The city and search pages show different details, and individual properties may omit optional data. Check sourcePage. An empty check-in time, room list, or description means it was not on the accessed page; it is not an estimate.

Can I paste a hotel URL? Direct hotel pages are outside the current input scope. Use a city or search-results URL and follow the url in the returned row to review that hotel on the website.

Integrations

Run the actor through its API, read the default dataset, and send the records to a database, spreadsheet, notification workflow, or reporting application. Scheduled runs and webhooks support recurring collection and completion alerts. Use propertyId as the stable key when merging snapshots, and retain sourcePage, priceBasis, dates, and currency when building comparisons over time.

Support

Open an issue on the Issues tab; we reply within 24h and add fields on request. Include a public city or search URL, the run ID, and the field you expected so the page behavior can be checked.