Reviews Scraper & AI Insights — 5 Sources in One
Pricing
from $2.00 / 1,000 reviews
All public reviews for a company or app across Trustpilot, G2, Capterra, the App Store and Google Play in one schema, with ratings, dates and company replies. Optional AI insights: themes, pain points, praise, feature requests and competitor mentions, in English or Arabic.
Each entry is { name, website, trustpilotDomain, g2Url, capterraUrl, appStoreId, appStoreUrl, googlePlayId, googlePlayUrl }. Only 'name' is required — 'website' is enough for Trustpilot, and app ids are looked up from the name when omitted.
Which review sources to read. G2 and Capterra are off by default: both block automated traffic aggressively, so selecting them usually returns a source error rather than reviews.
[ "trustpilot", "appstore", "googleplay"]Storefronts to read for the App Store and Google Play, as ISO-2 codes. Each storefront has its own reviews, so adding sa/ae/eg is how you get Gulf and Egyptian feedback.
[ "us", "gb", "sa", "ae", "eg"]Keep only reviews in these languages (ISO-639-1, e.g. en, ar). Leave empty for all.
Ignore reviews older than this. Reviews with no date are kept, so a source that omits dates is not silently dropped.
Keep only reviews at or above this rating, on the normalized 1-5 scale.
Keep only reviews at or below this rating. Set to 2 to collect complaints only.
Stop after this many reviews per target per source. This is the main control on cost.
Include the company's public reply to a review where the source publishes one.
Turn the collected reviews into themes, pain points, praise, feature requests, competitor mentions and suggested actions. Billed per batch of 10 reviews analyzed.
Language for the AI insights. 'ar' produces Modern Standard Arabic; review quotes stay in the language they were written in.
Return only what is new since the last run under the same tracking ID: new reviews, negative reviews, rating drops and newly added company replies. Off = return the reviews themselves.
Name of this watch. Required when 'Emit alerts only' is on. Snapshots are stored under this key in the 'review-intel-state' key-value store.
On the first run for a tracking ID, report every review as new. Off = the first run records a baseline silently so your first alert is a real one.
A new review at or below this rating also raises a separate negative_review alert — the one you route to a human.
Raise a rating_drop alert when a source's average rating falls by at least this much since the last run.
How many target x source pairs are fetched in parallel.