TripAdvisor Reviews Scraper — Hotel Guest Reviews & Ratings

A tripadvisor reviews scraper — a hotel reviews scraper for tripadvisor hotel reviews — that returns every review a hotel has as structured rows: tripadvisor ratings as 1–5 bubbles, the title and the full text, the six sub-ratings (value, rooms, location, cleanliness, service, sleep quality), the language it was written in and whether you are reading a machine translation, the stay date and trip type, the reviewer's profile, photos, helpful votes and the hotel's management response. It is a tripadvisor scraper for guest reviews built for review monitoring: give it hotels, get tripadvisor review data — hotel review data you can sort, filter and export.

Give it TripAdvisor hotel URLs, TripAdvisor location ids or hotel names. No login, no tripadvisor api key, no browser. It reads the same review pages a visitor sees — server-rendered, with the full review record inside — through a residential proxy and a browser-grade TLS fingerprint, so runs are fast and steady where plain scrapers are turned away. It bills per review row, and every filter (rating, trip type, date, text, translation) runs before billing, so a complaints feed pays only for complaints.

Last verified working: 2026-08-29.

What does the TripAdvisor reviews scraper do?

You give it hotels. It gives you their tripadvisor hotel reviews as rows, in the languages you choose.

• Every review, newest first. TripAdvisor lists reviews ten to a page, newest first; the actor walks the pages until it has what you asked for or the list ends. A 2,000-review hotel is 200 pages; maxReviewsPerHotel keeps that under control.
• The whole record. Bubbles, title, text, six sub-ratings, stay month, trip type (family, couples, solo, business, friends), the reviewer's name, handle, verified flag, contribution counts and hometown, attached photos, helpful votes, and the management response with its date. Each review also carries its own URL, so any row can be checked against the site.
• Thirty languages. TripAdvisor is not one site: tripadvisor.com serves English, tripadvisor.de German, tripadvisor.fr French, and so on — each domain shows reviews written in its language plus machine translations of the rest. languages picks the domains; each row says which language it was read in, which language it was written in, and whether it is a translation.
• A hotel row for context. Bubble rating, total review count, reviews per rating and per language, the six aspect averages, the city ranking ("#33 of 372 hotels in Lisbon"), address, coordinates, phone, website and amenities. Never charged.
• Filters that run before billing. maxRating for complaints feeds, tripTypes, sinceDate for schedules, requireText, includeTranslated. Filtered reviews are never charged.

TripAdvisor reviews by hotel: URL, location id or name

Three ways to point at a hotel, mixed freely:

• URLs — any TripAdvisor hotel page, on any domain: https://www.tripadvisor.com/Hotel_Review-g189158-d4509998-Reviews-Memmo_Alfama_Hotel-Lisbon.html. Only the numeric d4509998 matters; TripAdvisor redirects a wrong geo id or slug to the canonical page.
• Location ids — 4509998 or d4509998, the number after d in any TripAdvisor URL. What an earlier run's location_id gives you; the fastest path because nothing needs resolving.
• Names — "Memmo Alfama Lisbon". Resolved through TripAdvisor's own typeahead. Its answer must share at least half of the words you typed; a name TripAdvisor does not know comes back as an error row instead of a random hotel, and the resolved name is in the hotel and status rows so you can check the match.

A URL, its id and its name are one hotel: the actor resolves first, then dedupes, so the reviews are read and billed once, and the extra targets are reported as duplicates.

TripAdvisor negative reviews without paying for the happy ones

maxRating: 2 keeps only one- and two-bubble reviews; maxRating: 3 adds the middling ones. The rejected reviews are never billed. Because TripAdvisor serves reviews newest first with no rating sort, the pages are still read (each page is about a megabyte of residential proxy traffic, which is the actor's own cost, not yours), so pair a complaints feed with a sensible maxReviewsPerHotel or a sinceDate. A hotel with 1,599 five-bubble reviews and 30 low ones delivers and bills 30 rows out of the pages it read.

Add requireText: true to skip rating-only reviews, tripTypes: ["business"] to hear only from business travellers, and includeTranslated: false to keep only reviews written in the language you are reading.

Hotel review data for revenue, operations and analysts

• Hotel reputation management. Schedule one run a day with sinceDate: "2 days" for your properties and a comp set. Pages come newest first, so the run stops after the first page that predates the cut — two or three pages per hotel per day — and every new review lands in a dataset you can route to Slack, a sheet or a ticket.
• Tripadvisor competitor reviews. The hotel row's rating_counts, reviews_by_language and sub_ratings describe a hotel in one line; the review rows explain the numbers. Ten hotels, one run, one table.
• Voice of the customer. Six sub-ratings per review plus free text is the input every hotel sentiment analysis wants: the sub-ratings tell you what slipped (rooms 2, service 5), the text tells you why.
• Multi-language listening. A Lisbon hotel has 1,393 English reviews, 220 French, 141 German and 74 Portuguese. Read the French and German sites to hear those guests in their own words, with includeTranslated: false so each review arrives once.

Input

Field	What it does
startUrls	TripAdvisor hotel URLs on any domain; only the d<id> matters.
locationIds	TripAdvisor location ids, 4509998 or d4509998.
hotelNames	Hotel names, ideally with the city; resolved through TripAdvisor's typeahead with a word-overlap guard.
languages	Which sites to read, one job per hotel and language: en (tripadvisor.com), en-gb, de, fr, es, it, pt (tripadvisor.com.br), nl, ja, zh, ru, sv, da, no, fi, pl, tr, el, ko, th, vi, id, cs, hu, he, en-au, en-ca, en-in, en-ie, en-sg. Default ["en"].
maxReviewsPerHotel	0 = every review the site lists in that language; N = the newest N. Default 100. The main cost control.
maxRating	0 = all; 1–5 = keep reviews with that many bubbles or fewer. Filtered reviews are never charged.
tripTypes	Keep only family, couples, solo, business, friends. Empty = all.
sinceDate	Keep reviews published on or after this date — 2026-06-01 or relative ("30 days", "2 weeks"). Paging stops once a whole page predates it.
requireText	Drop rating-only reviews before billing.
includeTranslated	Include machine-translated reviews (default true). Off = only reviews written in the site's language.
includeHotelRow	Emit the free hotel context row per hotel and language (default true).
sessions, perIp	Residential proxy sessions in parallel (default 2) and the pace per session (default 1 request/s).
proxyConfiguration	Apify Proxy, RESIDENTIAL group — required, and the default.

Example: a scheduled complaints feed

Every morning, every new one- or two-bubble review of five properties, in English and German, with text:

{
  "locationIds": ["4509998", "228423"],
  "hotelNames": ["The Lumiares Lisbon", "Bairro Alto Hotel Lisbon", "Memmo Príncipe Real"],
  "languages": ["en", "de"],
  "sinceDate": "2 days",
  "maxRating": 2,
  "requireText": true,
  "maxReviewsPerHotel": 0
}

Most days this reads two pages per hotel and language and bills nothing, because nothing new was under three bubbles. When something lands, it is in the dataset before the morning meeting.

Example: the full French corpus of a competitor set

{
  "startUrls": [
    "https://www.tripadvisor.com/Hotel_Review-g189158-d4509998-Reviews-Memmo_Alfama_Hotel-Lisbon.html",
    "https://www.tripadvisor.com/Hotel_Review-g189158-d190364-Reviews-x.html"
  ],
  "languages": ["fr"],
  "includeTranslated": false,
  "maxReviewsPerHotel": 0
}

The .com URLs are fine even though the run reads tripadvisor.fr — the id is what counts. includeTranslated: false means each row is a review a French-speaking guest actually wrote.

Example: what business travellers say, by id

{
  "locationIds": ["d4509998"],
  "tripTypes": ["business"],
  "maxReviewsPerHotel": 300,
  "includeHotelRow": false
}

Output

Three row types share one dataset, told apart by type.

review — one per review, charged:

{
  "type": "review",
  "target": "https://www.tripadvisor.com/Hotel_Review-g189158-d4509998-Reviews-Memmo_Alfama_Hotel-Lisbon_Lisbon_District_Central_Portugal.html",
  "review_id": "1075128384",
  "location_id": "4509998",
  "hotel_name": "Memmo Alfama Hotel",
  "language": "en",
  "original_language": "en",
  "translated": false,
  "title": "A Gem in the Heart of Alfama",
  "text": "This place is a gem!\nOn a quiet side street just off the route of the famous tram Electrico #28, this is a gorgeous haven in the heart of Alfama. Our room had a view out over the Tagus River…",
  "rating": 5,
  "sub_ratings": {"value": 4, "rooms": 4, "location": 5, "cleanliness": 5, "service": 5, "sleep_quality": 5},
  "review_date": "2026-08-27",
  "created_date": "2026-08-27",
  "stay_date": "2026-08-31",
  "trip_type": "COUPLES",
  "helpful_votes": 0,
  "reviewer_name": "Frank S",
  "reviewer_username": "992franks",
  "reviewer_verified": false,
  "reviewer_contributions": 391,
  "reviewer_helpful_votes": 169,
  "reviewer_hometown": "Brooklyn, New York",
  "photos": [],
  "photo_count": 0,
  "response": "Dear Frank, thank you so much for your incredible review…",
  "response_date": null,
  "response_id": "974781822",
  "url": "https://www.tripadvisor.com/ShowUserReviews-g189158-d4509998-r1075128384-Memmo_Alfama_Hotel-Lisbon_Lisbon_District_Central_Portugal.html",
  "hotel_url": "https://www.tripadvisor.com/Hotel_Review-g189158-d4509998-Reviews-Memmo_Alfama_Hotel-Lisbon_Lisbon_District_Central_Portugal.html",
  "fetched_at": "2026-08-29T15:20:11+00:00"
}

rating is 1–5 bubbles. sub_ratings carries only the aspects the reviewer rated. language is the site the row was read on; original_language is what the guest wrote in; translated is true when the text is TripAdvisor's machine translation. stay_date is the month of the stay as TripAdvisor records it. trip_type is one of FAMILY, COUPLES, SOLO, BUSINESS, FRIENDS or null. response is the management response text, with response_date and response_id when TripAdvisor exposes them. photos are 1,200-pixel renditions.

hotel — one per hotel and language, free: location_id, geo_id, hotel_name, url, rating (the hotel's bubble rating, e.g. 4.7), review_count (all languages), rating_counts ({"1": 13, "2": 17, "3": 68, "4": 280, "5": 1599}), reviews_by_language ({"en": 1393, "fr": 220, "de": 141, …}), sub_ratings (six aspect averages), ranking ("#33 of 372 hotels in Lisbon") with ranking_position and ranking_of, address, city, postal_code, country_code, latitude, longitude, phone, website, price_range, amenities, reviews_listed (what the site lists in this language, translations included), reviews_fetched, fetched_at.

status — one per hotel and language plus one per duplicate or failed target, free: target, location_id, hotel_name, language, status (ok, no_reviews, not_found, duplicate, error), reviews delivered, filtered (dropped by your filters, never charged), pages read, total listed, duplicates, error, fetched_at.

The run's key-value store holds a SUMMARY record: hotels, jobs, reviews, filtered, duplicates, pages, not_found, errors, what was charged, whether the spending limit was hit, HTTP stats (including how many DataDome refusals were rotated past) and the run time.

How much does it cost?

$0.005 per review row. Hotel rows, status rows, filtered reviews, unknown ids, duplicate targets and reviews the spending limit refused are free. Every review a hotel has on the English site — say 1,456 — is $7.28; a daily complaints feed that finds two complaints a week costs cents a month. Set maxReviewsPerHotel and sinceDate to keep runs the size you want, and the run's maxTotalChargeUsd as a hard ceiling. Residential proxy traffic (about a megabyte per ten reviews) is included in the price.

TripAdvisor reviews scraper in Python, JavaScript, curl, n8n, Make or an AI agent

Python

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("kestrel/tripadvisor-reviews-scraper").call(run_input={
    "locationIds": ["4509998"],
    "languages": ["en", "fr"],
    "maxRating": 3,
    "requireText": True,
    "maxReviewsPerHotel": 500,
})
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    if row["type"] == "review":
        print(row["language"], row["rating"], row["review_date"], row["title"], "-", (row["response"] or "")[:40])

JavaScript

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_APIFY_TOKEN>' });
const run = await client.actor('kestrel/tripadvisor-reviews-scraper').call({
    hotelNames: ['Memmo Alfama Lisbon', 'The Lumiares Lisbon'],
    sinceDate: '30 days',
    includeTranslated: false,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
const reviews = items.filter(r => r.type === 'review');
console.log(reviews.length, 'reviews;', reviews.filter(r => r.response).length, 'answered by the hotel');

curl

curl -X POST "https://api.apify.com/v2/acts/kestrel~tripadvisor-reviews-scraper/run-sync-get-dataset-items?token=<YOUR_APIFY_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"startUrls": ["https://www.tripadvisor.com/Hotel_Review-g189158-d4509998-Reviews-Memmo_Alfama_Hotel-Lisbon.html"], "maxReviewsPerHotel": 50}'

n8n and Make. Use the Apify node (or a plain HTTP Request node calling run-sync-get-dataset-items as above), then a filter on type = review and whatever you want to do with the rows — a Slack alert for rating <= 2, a Google Sheet, a CRM ticket. The sinceDate field takes relative dates so the schedule never goes stale. Ready-made review-alert workflows for the sister actors are in github.com/mtedj/kestrel-actors-examples.

MCP and AI agents. The actor is on Apify's MCP server, so a Claude, Cursor or custom agent can ask for "the latest negative TripAdvisor reviews of hotel X" and get the rows; the input schema tells the agent what every field does.

Is it legal to scrape TripAdvisor reviews?

The actor reads public review pages that any visitor sees without logging in and returns what is on them; it stores nothing beyond the rows in your dataset. Reviews are user-generated content that reviewers published under a public name or handle: if you build a product on top, you are the data controller for whatever you keep, so retain what you need (for most review monitoring that is the text, rating and date), respect deletion, and read TripAdvisor's terms before scraping at scale — they restrict automated access, and the responsibility for how the data is used is yours. Do not use it to republish reviews wholesale; do use it to understand what guests say about a property.

Limits and honest notes

• Newest first only. TripAdvisor's pages sort by date and expose no rating or helpfulness sort, so maxRating reads pages and filters them. A complaints feed on a huge hotel is best combined with sinceDate.
• Language counts differ from the total. review_count on the hotel row is the all-language total (1,977); reviews_listed and total are what the site you read lists in its language, translations included (1,456 on the English site, 595 on the German site). Native-language counts are in reviews_by_language.
• A translation is a separate row per language. With languages: ["en", "de"] a German-written review appears once from tripadvisor.de (original) and once from tripadvisor.com (machine-translated). includeTranslated: false keeps each review exactly once.
• Residential proxy is not optional. TripAdvisor's bot gate refuses every datacenter range; the actor also needs a Firefox-grade TLS fingerprint, which it carries itself. A refused page is rotated and retried, and after six refusals the hotel is reported as an error — never as "no reviews".
• Pages are heavy. About one megabyte per ten reviews, paced at one page per second per session. Two sessions read roughly 20 reviews a second.
• Hotels only. Restaurants, attractions and vacation rentals have different page shapes; give the actor Hotel_Review pages.

FAQ

Does it need a TripAdvisor API key or login?

No. There is no public tripadvisor api for reading reviews (the Content API is for partners and does not return review text); the actor reads the public review pages, which carry the full record.

Can I download TripAdvisor reviews as CSV or Excel?

Yes. Every run writes an Apify dataset; open it in the Console and export tripadvisor reviews CSV, Excel, JSON or XML, or pull it through the API. The reviews and complaints views are ready-made column layouts.

How do I get only TripAdvisor negative reviews?

Set maxRating to 2 (or 3). Reviews above it are dropped before billing. Add requireText: true so every row has something to read, and sinceDate for a rolling window.

Can I get TripAdvisor reviews by location id, URL or name?

Any of the three. Ids are fastest; URLs need no lookup either (only the d<id> in them matters); names go through TripAdvisor's typeahead with a guard against wrong matches.

How many reviews can one hotel return?

All of them, per language: maxReviewsPerHotel: 0 walks every page the site lists. Large hotels have thousands; a sinceDate or a cap keeps a scheduled run small.

Can I get TripAdvisor reviews in multiple languages, and are they translated?

You choose the sites with languages; each site shows reviews written in its language plus machine translations of the others, marked translated: true with the original_language. Untick includeTranslated to keep originals only.

Does it include the TripAdvisor management response?

Yes — the tripadvisor management response is in response, with response_date and response_id when TripAdvisor exposes them.

Can I scrape TripAdvisor reviews without an API?

That is what this is — tripadvisor reviews without API access: the review pages are read directly, no GraphQL, no tokens, no browser. If TripAdvisor changes its pages the actor is fixed, and a daily canary run catches breakage within a day.

Can I use it for hotel sentiment analysis?

Yes — the six sub-ratings, the free text and the trip type per review are the inputs sentiment work needs, and the original_language lets you route each text to the right model.

What do bulk hotel reviews cost?

$0.005 per review, nothing else. Ten hotels × 500 newest reviews = 5,000 rows = $25; the daily feed that follows costs only what is new.

Review monitoring across a portfolio: TripAdvisor guest feedback every morning

Put every property and its comp set in locationIds, schedule the actor daily with sinceDate: "2 days" and maxReviewsPerHotel: 0, and dedupe on review_id + language downstream. Each run reads two or three pages per hotel and language, delivers the new reviews, and its status rows tell you how many were filtered, how many pages were read and whether any hotel failed.

Guest feedback data that keeps its structure

Rows never change shape: every review row carries every field (null where TripAdvisor has nothing), so a sheet or a table built on the first run keeps working. sub_ratings and photos are nested where the data is naturally nested and flat everywhere else.

What this does not do

It does not write reviews, vote, flag or contact reviewers; it does not read restaurants or attractions; it does not sort by rating (TripAdvisor does not expose that on its pages); and it does not bypass TripAdvisor's residential-only gate with anything other than a residential proxy.

Choosing languages

Read the site your guests use. For an international hotel that is usually en plus the two or three languages with the highest reviews_by_language counts on the hotel row; with includeTranslated: false each review arrives once, in the words the guest wrote.

Related scrapers

Hotel reviews live on more than one site. These read the others, and they share the same row conventions, the same pay-per-delivered-row billing and the same scheduling story:

• Despegar Hotel Ratings Scraper — Latin America's largest OTA: guest score, six category scores and the site's own summary of what guests say, per property.
• TripAdvisor Restaurant Reviews Scraper — the same engine on restaurants and attractions, with the food/service/value/atmosphere sub-ratings and the dish keywords a hotel page does not carry.
• Hostelworld Reviews Scraper — hostel guest reviews with eight sub-ratings, traveller group and trip type, and the owner's reply.
• Booking.com Reviews Scraper — Booking.com guest reviews with liked and disliked text split, traveller type and the property's reply.
• Agoda Reviews Scraper — Agoda hotel reviews with separate positives and negatives.
• Trip.com Reviews Scraper — Trip.com and Ctrip reviews with four sub-scores and the Chinese-language corpus.
• HRS Reviews Scraper — HRS business-travel reviews with twelve category votes.
• Google Hotels Prices Scraper — what the same hotels charge, per booking source and stay.

All of them bill per delivered row, never charge for rows a filter or a spending limit removed, and write an Apify dataset you can export to CSV, Excel or JSON.