Hotel Reviews Monitor - New Agoda, Trip.com, Kurzurlaub Reviews

For hotel revenue managers and review-alert tools that check each property daily: the reviews not returned before, from Agoda, Kurzurlaub.de and Trip.com (the about 10 on its hotel page), with rating, date, text, reviewer, room or package, category scores and the hotel's reply. In production runs on 2026-09-23, one property per run took 2 to 29 seconds and a check with nothing new returned one free row. Put one property on an hourly or daily schedule and you get a row only when a guest writes something, with no start date to track and no diffing on your side.

Unofficial. Public data only.

Export as JSON, CSV or Excel, or read the dataset from the API after each scheduled run.

What you get

• Only new reviews. Every run remembers which reviews it has returned for each property. The next run returns only the ones it has not returned before. You never type a start date.
• One small check per property. Built to be run often: one property per run, usually one request and a few seconds (Agoda hotels with reviews in many languages: up to about 40 requests and 30 seconds). Put each property on its own schedule, or list up to 20 in one run.
• Three sites, one row format. Agoda (including the Booking.com and Priceline.com reviews Agoda displays), Trip.com and Kurzurlaub.de come back with the same columns, and source and reviewSite say where each review is from.
• A clear answer every time. A property with nothing new, no reviews, a page that does not exist, a refused request or a spending limit each comes back as a free row that says why. A scheduled run never ends silently.

Measured on 2026-09-23 (Apify production, one property per run)

Site	Property	First check	Later check with nothing new	Requests per check	Platform cost per run
Kurzurlaub.de	Schloss Lieser (187 reviews)	4.5 - 6.7 s, 10 rows (9.0 s for 25)	1.6 - 4.1 s, 1 free row	1 (+1 when the page URL is new)	$0.00009 - $0.00048
Trip.com	5L Hotel, Beijing (2,324 reviews)	4.6 - 6.7 s, 10 rows	3.0 - 10.6 s, 1 free row	1	$0.00005 - $0.00031
Agoda	Marine Hotel, Dublin (1,389 reviews incl. Booking.com)	18.6 - 20.8 s, 10 rows	4.7 - 14.0 s, 1 free row	1 - 40	$0.00005 - $0.00049

First checks of Imperial Hotel Tokyo and Sheraton Grand Taipei on Agoda took 29 s and 22 s; Kurzurlaub.de and Trip.com first checks of five other hotels took 3-5 s. Run times include the container start. Agoda needs more requests for hotels whose reviews are spread over many languages (see "How new reviews are found").

Input

Field	What it does
propertyUrls	Hotel page URLs from Agoda (https://www.agoda.com/marine-hotel/hotel/dublin-ie.html), Trip.com (https://www.trip.com/hotels/beijing-hotel-detail-375126/) or Kurzurlaub.de (https://www.kurzurlaub.de/hotel/mosel_saar/lieser/hotel-schloss-lieser.html, or one of the hotel's package pages). Up to 20 per run.
maxReviewsFirstRun	The first time a property is checked, return this many of its newest reviews and only remember the rest. 0 = return none and start watching from now. Empty = 10. A first check reads at most 50 reviews on Kurzurlaub.de and about 10 on Trip.com, so a larger number returns no more there. Later runs return every new review.
includeOtherSites	Agoda only. On or empty = also return the Booking.com and Priceline.com reviews Agoda shows. Off = only reviews written on Agoda. Switching it starts a new watch for Agoda properties (the next run is a first check again).
watchName	Optional. Runs with the same watch name share what has already been returned. Give two schedules that watch the same property different names so each gets every new review.
resetMonitoringState	Forget what this watch remembered for these properties at the start of this run, so this run is a first check again. Turn it off again for scheduled runs.

Output

{
  "source": "trip.com",
  "status": "ok",
  "propertyId": "375126",
  "propertyName": "5L Hotel",
  "propertyUrl": "https://www.trip.com/hotels/beijing-hotel-detail-375126/5l-hotel/",
  "inputUrl": "https://www.trip.com/hotels/beijing-hotel-detail-375126/5l-hotel/",
  "reviewId": "2106132610",
  "reviewSite": "Trip.com",
  "rating": 5.5,
  "ratingScale": 10,
  "ratingText": null,
  "reviewDate": "2026-09-17",
  "reviewDateTime": "2026-09-17T15:21:25",
  "stayDate": "2026-09",
  "title": null,
  "text": "The room was very old. TV wasn't working well. There was no telephone in the room, so I could only reached front desk by using my cell phone to call. Toilet bowl flushing system was faulty.",
  "positives": null,
  "negatives": null,
  "language": "en",
  "translatedTitle": null,
  "translatedText": null,
  "translatedLanguage": null,
  "reviewerName": "YSTT",
  "reviewerCountry": "Singapore",
  "travelerType": "Traveling with friends",
  "roomType": "Apartment Of Twin Beds",
  "nightsStayed": null,
  "packageName": null,
  "categoryRatings": { "location": 8, "amenities": 4, "service": 6, "cleanliness": 4 },
  "helpfulVotes": 0,
  "hotelResponse": "Dear Guest, thank you for sharing your feedback with us. We’re truly sorry that your stay did not meet your expectations. ...",
  "hotelResponseDate": "2026-09-17",
  "photoUrls": [],
  "positionOnPage": 1,
  "isFirstCheck": true,
  "propertyReviewCount": 2324,
  "propertyRating": 9,
  "propertyRatingScale": 10,
  "checkedAt": "2026-09-23T11:03:45.616Z"
}

• rating is on the scale in ratingScale: 10 for Agoda and Trip.com, 6 for Kurzurlaub.de (its own 1-6 points).
• reviewDate is the date the site gives for the review. stayDate is the stay (Agoda: check-in day; Trip.com: month only).
• isFirstCheck is true for rows from a property's first check (its newest reviews, not necessarily new ones) and false for reviews that are new since the watch started.
• propertyRating is not filled for Agoda (it would cost an extra request on every run).
• positionOnPage is only filled for Trip.com.
• Columns a site does not have are null, never guessed: Kurzurlaub.de has no review title, reviewer country or hotel reply; Trip.com has no title or pros and cons; Booking.com reviews shown on Agoda often have only pros and cons, no text.

Free rows have a status other than ok and a note that says why: no-new-reviews, watch-started (first check with maxReviewsFirstRun 0), no-reviews, not-found, blocked, unreadable, bad-input, budget-reached.

How new reviews are found

• Kurzurlaub.de lists a hotel's reviews newest first, 10 per page. A check reads page 1 and goes on only while every review on the page is new. Kurzurlaub.de reviews have no ID of their own, so the ID is made from the hotel, date, reviewer first name and package; two reviews by the same first name on the same day for the same package count as one.
• Agoda sorts newest first, but within each language, so a new Spanish review can sit far down the list. A check reads the first 100 reviews of each site and then the newest 20 of each language that did not fit. Agoda also serves two slightly different copies of its review list to different visitors; a review that shows up for the first time but is dated more than 7 days before the watch started is remembered and not returned (the free row says how many).
• Trip.com shows about 10 reviews on the hotel page in its own "Most relevant" order, not newest first, and this Actor reads only that page. A review is returned when it first appears among them (positionOnPage says where), and the same 7-day rule applies. New Trip.com reviews that never reach those 10 are not returned.
• The first check returns up to maxReviewsFirstRun of the newest reviews. If a first check could not finish (for example it reached the spending limit), the next run finishes it and also returns every review dated from 2 days before that first check on.
• A check reads a limited number of pages (Kurzurlaub.de 5 pages; Agoda 3 pages of 20 per language). If every review on the last page is new, older new reviews beyond it may never be returned, and a free unreadable row says so; check busy properties more often.
• Hostelworld is not supported: its review service refuses Apify's network, and its property page shows only the 3 latest reviews without IDs.

Scheduling tips

• One property per run keeps each run to seconds and makes each property's result its own dataset.
• Do not put the same property in two overlapping schedules with the same watchName. Apify's key-value store has no atomic update, so two runs finishing at the same moment can return the same new review twice. Different watchNames never interfere.
• The run start fee is charged even when nothing is new; that is what pays for the check.

Pricing

Pay per event:

• Run start: $0.02 per run (Bronze $0.0182, Silver $0.0164, Gold $0.0146), charged once before the first request, whether or not a new review is found, and also when the site refuses the request. A run whose input cannot be used, or whose maximum total charge has no room for the start fee plus one review row, requests nothing and is charged nothing. A run restarted or moved by Apify is charged once.
• New review returned: $5.00 per 1,000 (Bronze $4.50, Silver $4.00, Gold $3.50). Only rows with status ok are charged; every free row explains itself.

A daily check of one property with nothing new costs $0.02. A check that finds 3 new reviews costs $0.02 + $0.015.

Limits

• Up to 20 properties per run.
• No proxy is used and site checks are never bypassed. A refused request (HTTP 403 or 429, or a check page) stops that site for the rest of the run with a free blocked row; nothing is remembered for that property, so the next run reads it again.
• Reviewer names are the display names the sites show publicly. Email addresses and long phone numbers written inside review text are masked.