# Changelog of TripAdvisor Reviews Scraper — Search & Monitor (`zenomastro/tripadvisor-reviews-reliable`) Actor

- **URL**: https://apify.com/zenomastro/tripadvisor-reviews-reliable/changelog.md
- **Full Actor documentation**: https://apify.com/zenomastro/tripadvisor-reviews-reliable.md

## Changelog

### 2026-10-14

- Free rating distribution: the place summary row now has the exact 1-5 rating counts of all the place's reviews (`ratingCounts`) and their average (`averageRating`), from TripAdvisor's own counts. Airline pages keep the total count only.
- Richer review insights: average sub-ratings (`subratingAverages`), traveler-type and language distributions, and the share of reviews with photos.
- New review fields: `photoCount` (TripAdvisor's photo count), `reviewTips` (reviewer tips), `ownerReplyAuthor` and `ownerReplyRole` (who answered), `stayMonth` (month of the stay or visit, YYYY-MM). Existing fields are unchanged.
- New filter: months of the stay or visit (`visitMonths`), e.g. June to August for the summer season.
- New input `startUrls`: place URLs with your own `userData`, returned as `customData` on every row of the place; rows of another TripAdvisor dataset with a `url` or `webUrl` field work too.
- New option `includeReviewerInfo`: turn it off to collect reviews without reviewers' personal data.
- Links to a single review (ShowUserReviews) are accepted and collect the reviews of their place.
- Payloads written for other TripAdvisor review scrapers work as they are: `maxItemsPerQuery`, `maxReviews`, `reviewsPerPlace`, `lastReviewDate`, `since`, `disableMachineTranslations` and `scrapeReviewerInfo` are accepted.
- Reliability: the run stops shortly before its timeout instead of timing out, delivers everything collected and says how far it got; a platform migration or a resurrected run continues where it stopped without delivering or charging any review twice; progress and the review monitor state are saved during the run (periodically and on a graceful abort), not only at its end.
- Faster delivery: reviews are written and charged in batches of 50 instead of one by one.
- Progress in the status message while a large place is being read.
- Input form reorganised into sections (places, filters, output, monitoring, advanced) with clearer descriptions; two new Output views: review details, and place summaries and insights.

### 2026-10-03

- Fixed: TripAdvisor retired the review query; migrated to the new review API. Hotels, restaurants and attractions return reviews again (airlines were not affected). Output field names are unchanged.
- Rating and traveler-type filters are now applied by TripAdvisor itself, so filtered runs need far fewer requests. The place summary still reports the place's total review count.
- New review fields: `publishedAt` (exact publication time) and `subratings` (sub-ratings as a label → score map). `reviewerHometown` shows the place name again, and airline reviews now include the airline name. Place summaries return `accommodationCategory` as `null`, like review rows.
- Changed at the source: TripAdvisor now exposes one review date, so `createdDate` equals `publishedDate`; `stayDate` keeps the last day of the stay month; `accommodationCategory` and `publishPlatform` are no longer provided (`null`).
- `machineTranslation` off still returns every review in its original language (read from TripAdvisor's per-language lists, which can take a few extra requests).
- Review languages return only reviews written in those languages, in their original text, whatever the translation setting (TripAdvisor also lists machine translations of other reviews under each language; they are skipped and not charged). Each review comes once, whatever the order of the languages.
- Very deep pages: when TripAdvisor stops serving a place part-way (seen around 10,000 reviews deep), the reviews collected so far are delivered and the place summary says so (`scanTruncated`, `scanTruncatedReason`) instead of the place failing.
- The same place given as a URL and as a location ID (or found again by search) is scraped once, and a review is never delivered or charged twice in a run.
- Review languages accept spellings such as `zh-TW`, `zh-CN`, `he` and `id`; a language TripAdvisor cannot filter on is skipped with a warning instead of failing the place.
- Lower run start fee: $0.0001 per GB of run memory (was higher), so short runs and AI-agent calls cost less.
- Clearer title and description, and README cleaned up (no duplicated sections).
- Input form: the essential fields come first; optional settings are grouped under "Advanced settings".

### 2026-10-02

- New `machineTranslation` option: keep reviews in their original language (default) or get TripAdvisor's English machine translation.
- Date filters stop paging as soon as a page is older than the requested window, so recent-only runs finish faster.
- Monitoring fix: only reviews actually delivered to your dataset are remembered as seen. If a run stops at your maximum charge limit, the undelivered reviews come in the next run.
- README: pricing section matches the live price (platform usage included, subscriber discounts).

### 2026-10-01

- Fixed: search by place name works again after TripAdvisor retired its search API. Place URLs keep working directly and are now the prefilled example.
- New Store icon.
- The monitor input is now called monitorName (monitorKey still works for existing tasks and schedules).
- Subscriber discounts on paid Apify plans: Bronze −10%, Silver −15%, Gold and higher −20% per result.
- Simpler pricing: Apify platform usage is now included in the per-result price, so you only pay per result plus a small start fee.
- Prefilled example input, so the first run works with one click.
- New Output tab with a results table and a JSON link for APIs and AI agents.
- README section on using the Actor from AI agents (Apify MCP server) and the REST API.
- Safer memory limits to keep runs fast and costs predictable.
- Residential and SERP proxy groups are replaced by the included datacenter proxy (your own proxy URLs still work), so the price stays all-inclusive.
- At the end of a run, the status message now links to the reviews and issues pages, so you can rate the Actor (1–5 ★) or suggest improvements.