| Review scrape One Product Hunt review, charged only when the row is actually written and only for reviews that are genuinely new. Product Hunt repeats a fixed block of top reviews on every page - measured, page two returned 18 review objects of which only 8 were new - and this Actor deduplicates by review id, so you are never billed twice for the same review. Measured over 260 real reviews, every charged row carries an overall rating, the reviewer's name and username, and a real UTC timestamp; 100% carry written text. Product Hunt has two review formats and the split is honest here: 62% are quick reviews with one written paragraph, and 22% are fully structured with four extra ratings (ease of use, reliability, value for money, customization) and prose split into what the reviewer liked, what they did not, and why they chose this over alternatives. Set reviewDetail to 'structured' to be charged only for the latter - filtered rows are never written and never charged. Per-product coverage rows are never charged. Error rows are never charged. A request that came back challenged and was retried is never charged. | | |
| Product process Charged once per product that produced a coverage row, INCLUDING a product with genuinely zero reviews, because 'this product has no reviews yet' is a real answer that cost a real request. The row it pays for reports how many reviews were collected against the total Product Hunt itself reports, the product's aggregate rating, its own AI-generated pro and con tags with the product-level view, how many pages were read, and why the run stopped. A product that could not be scraped at all writes an error row instead and is NOT charged. | | |