A Better, Cheaper Rainforest API Alternative

History is included in every record, wider series too, as the reference states. Figures assume the allowance is spent on one operation.

• One mode — every request is a scrape Rainforest re-scrapes the listing on every call, whether or not anything has changed, so you wait for a fetch even when the answer is identical to an hour ago. You choose per call instead: mode=cache answers from the warehouse in under 100 ms, mode=live forces a fresh scrape, and mode=auto uses the cache while it is fresh enough and scrapes only when it is not. Rainforest keeps no warehouse, so there is no cached path and no history to fall back on.
• Data that doesn't match the page Users report responses missing the buybox, description, or price while all three are visibly on the live page — a parse-quality lottery on every request. Our records come from a continuously validated pipeline with dedup and variant collapsing, and every record is timestamped so you know exactly what you're getting.
• Credits that run out fast Their entry plan is 500 requests a month, and past your allotment every extra request is billed as overage — $0.06 a credit on Hobbyist, from their own pricing page. Ours are published per call, the full offer list comes inside the half-credit record rather than costing a credit per page, and a miss or an empty result is never billed at all.
• Support tickets that age in months Public reviews describe tickets open for four to six months. We're a small engineering team; you talk to the people who run the pipeline.

• Warehouse or live scrape — your call mode=cache answers in under 100ms for half a credit, batch or single call. mode=live scrapes the listing on demand when it has to be current, for one credit. Failed scrapes are never billed.
• Barcode lookup Pass any UPC, EAN, GTIN or ISBN-13 instead of an ASIN — /v1/product?gtin=, /v1/offers?gtin= or a bare barcode as the /v1/shopping query. The product endpoints find any barcode, inside the marketplace you asked for: from the listings we observe, the same product on another Amazon marketplace, or Amazon’s own search (organic results only), and every answer says how and whether the barcode is printed on that listing. A barcode that maps to several ASINs returns every match; one nothing resolves is never billed.
• Current snapshot Title, brand, price and currency, the main image, BSR — plus the best category rank for the products that carry no overall rank — rating, review count, stock state, UPC, parent ASIN, the canonical product URL, and brand stats. Amazon's own "bought in past month" figure comes too, where it publishes one.
• Every seller, in the same call The full live offer list is standard, not a separate endpoint: every seller with name and ID, price and shipping cost, condition, FBA and Prime flags, positive-feedback percentage, lifetime feedback count, and buy-box days. New and used offers are split out, each with its own lowest offer and count. Gathering offers is the slow part of any live scrape, which is why most APIs bill them separately.
• Price and stock history The full price series plus price-change and stock-change events, in the base call.
• Rank history, per category Add bsr_history=true for daily rank in every category the product sits in. Included in the record, like the seller history below.
• Seller history — who held the buy box, and when Add offer_history=true for a buy-box timeline plus a per-seller price series, each with first seen, last seen and buy-box days. Included in the record too.
• Seller profiles in one call Rainforest's seller_profile, seller_feedback and seller_products map onto /v1/seller and /v1/seller/products — rating, feedback per period, histogram, five recent feedback entries (date, rating, struck; never a buyer's name or text) and the business block together, from our stored record at half a credit or live for one. Add history=true for the seller's rating-change and storefront-size series since we first observed it.
• Derived intelligence — what a scraper cannot return A page fetch returns one page. /v1/brand, /v1/brand/network, /v1/seller/brands, /v1/category and the ranked lists return what our buy-box observations add up to across products, sellers and months: how many sellers win a brand’s buy box and how that is trending, the brands connected by shared sellers (the seller map), every brand an operator was observed winning on, and the size, Amazon share, FBA share and concentration of a category. One credit per record or page of 25, lookups free, refreshed daily, never estimated.
• Fulfilment — AMZ, FBA or FBM Who actually ships it: the buy-box winner's fulfilment type, per-offer FBA and FBM flags, a count of each across the listing, and whether Amazon itself is selling or holding the box.
• Delivery days Read from Amazon's own delivery sentence at the time of the observation — earliest and latest days, the fastest Prime option where one is shown, and the promised date, for the buy box and for every offer.
• The page content itself Description, feature bullets, the full image gallery, category breadcrumbs, and variations with parent ASIN — plus featured customer reviews carrying author, star rating, title, body, verified-purchase flag, review date and helpful count. Each block is stamped with when the content was last captured.

Eleven Amazon marketplaces plus US Walmart, in one schema.