Ctrip / Trip.com Hotel Scraper

Scrape hotel listings from Trip.com (the English face of Ctrip): prices, ratings, reviews, location, star class, amenities, room types, and photos for any city and date range.

What you get per hotel

Field	Description
hotel_name	Hotel name
hotel_id	Trip.com internal hotel ID
url	Detail page URL
price	Lowest nightly rate seen
currency	Currency code (USD default)
original_price	Crossed-out / pre-discount price when present
rating	User rating (0-5 scale on Trip.com)
rating_label	"Excellent" / "Very good" / etc.
reviews_count	Total user reviews
stars	Star class (1-5)
address	Street / area description
city, district	Location breakdown
latitude, longitude	Map coordinates
image_url	Primary photo
amenities	List of facility names
room_types	Sample room types from card
tags	Promotional / property tags
distance_from_center	When Trip.com provides it
checkin_date, checkout_date, search_city	Echoed inputs
source	"api" (preferred) or "dom" (fallback)

Input

{
  "city": "Tokyo",
  "cityId": 228,
  "checkIn": "2026-06-01",
  "checkOut": "2026-06-03",
  "guests": 2,
  "rooms": 1,
  "maxResults": 50
}

cityId is strongly recommended for production. Trip.com's hotel list endpoint requires a numeric city ID — passing only a city name does not drive a search. The actor has a small built-in map of verified IDs (Tokyo 228, Singapore 73, Beijing 1, Shanghai 2, Madrid 357, Granada 717, New Delhi 495) and tries an in-browser autocomplete fallback for everything else, but the autocomplete resolver is not 100% reliable. To grab a cityId by hand, open https://www.trip.com/hotels/list?city=<id> for guesses and check the page title.

How it works

Trip.com renders its hotel list entirely client-side, behind a stealth-defence script (hotel-spider-defence) and a jigsaw CAPTCHA. Pure HTTP requests to the public fetchHotelList SOA endpoint are rejected without the dynamic testab token minted by the in-page JS.

The actor uses Camoufox (stealth Firefox) with Apify Residential proxy, performs a homepage warmup to settle session cookies, and then:

1. Attaches a page.on("response") listener before navigation, capturing every JSON response on /restapi/soa2/34951/fetchHotelList.
2. Loads /hotels/list?city=<id>&checkin=&checkout=&adult=&crn= and scrolls to trigger pagination / dynamic refresh XHRs.
3. Recursively walks each captured JSON for hotel records (matches dicts that look like a hotel: have hotelBasicInfo, or both hotelName and hotelId).
4. Falls back to anchor-based DOM extraction (a[href*="/hotels/detail/"] walked up to the smallest container with an image + meaningful text) if the API capture comes up short.

This is the same XHR-intercept pattern used by meesho-scraper and tatacliq-scraper in this repo.

Pricing (PPE)

Tier	Per result
FREE	$0.008
BRONZE	$0.006
SILVER	$0.005
GOLD	$0.004

Known limits

• Pagination is XHR-driven, not URL-driven. We rely on scrolling to fetch additional pages; 50-100 hotels per run is the comfortable ceiling. Larger maxResults may stall.
• CAPTCHA risk. Trip.com occasionally serves a jigsaw challenge to fresh proxy IPs. The actor retries up to 3 times with new browsers; persistent failures usually mean the proxy IP is currently flagged.
• Currency. All requests are pinned to USD via the curr URL param. Trip.com may still echo back rates that include local taxes/fees.
• City ID auto-resolve is best-effort. For ambiguous city names, supply cityId directly.