Traveloka Review & Sentiment Extractor
Pricing
from $1.40 / 1,000 results
Download Traveloka reviews in bulk from a hotel or Xperience URL: review text (original and translated), score, date, reviewer, photos, and aspect sentiment tags with counts. Ready for NLP. HTTP-only, no browser.
Traveloka hotel, Xperience activity, bus, airport-transfer or car-rental page URLs. The product family is detected from the URL path, so you can mix them freely.
Bare numeric Traveloka object IDs. These use the Product type setting below, since there is no URL to infer it from.
Which Traveloka product the IDs belong to. Only used for bare Object IDs and for URLs whose path cannot be classified.
How many reviews to download for each property or activity. Reviews arrive in pages, so this is also the stopping point for pagination.
Order Traveloka returns reviews in. Newest first is best for monitoring; lowest rating first is best for a complaint audit.
Keep only reviews in a rating band. Traveloka exposes bands, not exact score cut-offs.
Keep only reviews that have text, media, or both. Useful when you need photos, or when you want to drop score-only ratings before running NLP.
Keep only reviews carrying these aspect tags, for example GOOD_STAFF_FRIENDLINESS or BAD_CLEANLINESS. Run once without this to see which tags an object has, listed in the summary record under ratingTags.
Emit one OBJECT_SUMMARY row per property or activity holding the overall rating, total review count and aspect tags, alongside the individual reviews.
Collect the aspect tag aggregate (cleanliness, staff, location and so on) with counts, split into positive and negative tag IDs.
How many reviews to ask for per API call. Twenty matches what Traveloka's own web client asks for.
Traveloka locale. This affects the language of aspect tag labels and rating labels, and which translation of a review text you receive.
Currency context for the session. Reviews carry no prices, so this rarely matters.
Seconds to pause between API calls. Raise it if you see repeated 429 or 5xx responses on large pulls.
