Tutorials, industry analysis, and best practices for ecommerce data extraction, API integration, and competitive intelligence.
/amazon-seller-review only rejects a blank sellerId; every other bad value, from a Buy Box placeholder to Amazon's own ID, reaches the seller lookup as a billed 404. A client-side check against the real seller ID shape and Amazon's own first-party IDs catches those before they cost a credit, tested against a fake HTTP layer with the real helper modules.
/amazon-product-statistics returns buyBoxPrice as an integer in cents, while /amazon-product-lookup and /amazon-product-offers return decimal currency units, and offers.buyBox.totalPrice is an unrounded float on top of that. A tested Python script normalises all 3 into one comparable price and flags a mismatch when the conversion is missed, the class of bug that makes a mixed-endpoint total 100 times too high.
Since 19 Sep 2026, /amazon-product-offers, /amazon-bestsellers and /amazon-deals answer a retryable billed 5xx instead of a misleading 200 with a zero price or an empty list. Update client retry logic to trust the 5xx, stop treating a real 0.0 or empty result as a failure, and drop the old 5-7 attempt loop.
Amazon marketplace currency codes, names, symbols and default postcodes for all 23 marketplaces, straight from /amazon-get-marketplaces: 17 currencies, a shared dollar sign on two of them and the one marketplace with no decimal places. Includes a tested Python script that builds a cached lookup table and formats invoice lines correctly.
Build a live countdown list of Amazon Lightning Deals with the deals API: filter on isLightningDeal, parse dealEndTime as UTC, drop deals with no end time or that already ended, and sort the rest into ending-soon, today and later buckets - tested Python, one credit per poll.
The category_id parameter on /amazon-deals is never validated: an unknown department slug is not rejected, it is passed straight to Amazon's filter and just returns fewer or no deals, billed like any other answer. Check category_id against the real category tree before you call, the same way /amazon-bestsellers is checked, because here there is no error to catch a typo after the fact.
A 400 is free on one SellerMagnet endpoint and billed on another, and a malformed ASIN costs a credit on lookup but not on statistics. This is the per-endpoint map of which errors cost a credit, which are refunded, and which are a free gate - with a small classifier you can run before deciding to retry.
Three myths about the SellerMagnet sales estimate, checked against Amazon's own documentation and the API's code: it is not an exact count, most ASINs never get the 'bought in past month' figure at all, and a low estimate usually means thin signal, not zero sales. Two tested scripts show how to tell which path produced a given number.
Calculate the Amazon referral fee on a live Buy Box price - percentage of the total price or a stated minimum, whichever is greater - then subtract it and your fulfilment fee from the current price to see the margin over your landed cost per ASIN, in Python.
Join thousands of developers using SellerMagnet API for real-time Amazon product intelligence. Start extracting data in minutes.
150 free API credits • No credit card required • Cancel anytime
Manage your cookie preferences. Strictly necessary cookies are always enabled.
These cookies are essential for the website to function.
These cookies help us understand how visitors interact with our website.
These cookies are used to track visitors and display relevant ads.
Tell us about your project and we'll get back to you within 24 hours.