Amazon Product Scraper — Zip-Code Accurate Prices

Scrape Amazon product data by ASIN, product URL, or search keyword across 15 marketplaces. 28 fields per product — real-time prices, Best Sellers Rank, rating distribution, specs, seller, demand signals, delivery estimates and Amazon's own review aspects — priced and dated for the zip code you choose, not whatever region the scraper happened to run from.

No Amazon API key. No Amazon login. No scraping infrastructure to maintain.

---

Why this one

Amazon shows different prices, delivery dates and availability depending on where the shopper is. Most scrapers ignore this and silently return whatever their proxy saw — so your "price history" is really a history of random data-centre locations.

This Actor sets a real delivery location first, then scrapes. And it bills that setup once per run, not once per product.

	Typical competitor	This Actor
Location-accurate pricing	rarely, or billed per result	yes, billed once per run
Charged for not-found items	often	never
Charged for blocked retries	often	never
Minimum charge	commonly applied	none
Missing price returned as	0	null (a 0 would mean "free")
Delivery estimate per zip	rarely	yes, deliveryDate moves with the zip
Duplicate rows billed	sometimes	never — deduplicated before charging
Warns when your proxy degrades results	no	yes, during the run and in RUN_SUMMARY

Worked example — 100 products with a zip code:

This Actor:  100 x $0.002  +  $0.02 once      = $0.22
Per-result location billing: 100 x $0.038     = $3.80

---

Quick start

Pick any one of three input modes. You can combine them.

By ASIN

{ "asins": ["B07CMS5Q6P", "B07L4BM851"] }

By URL

{ "urls": ["https://www.amazon.com/dp/B07CMS5Q6P"] }

By keyword search

{ "search": ["wireless gaming mouse"], "maxItemsPerStartUrl": 20 }

With location-accurate pricing

{ "asins": ["B07CMS5Q6P"], "zipCode": "90210" }

---

Input reference

Field	Type	Default	Description
asins	string[]	[]	Amazon ASINs, e.g. "B07CMS5Q6P".
urls	string[]	[]	Amazon product URLs. The ASIN is extracted automatically.
search	string[]	[]	Search keywords. Every matching product is scraped.
maxItemsPerStartUrl	integer	20	Max products collected per search keyword. The Actor paginates Amazon's result pages to reach it.
marketplace	string	"amazon.com"	One of the 15 marketplaces listed below.
zipCode	string	null	Delivery zip code for location-accurate prices and stock. Cannot be combined with countryCode.
countryCode	string	null	Delivery country as an alternative to zipCode. Cannot be combined with it.
proxyConfiguration	object	Apify proxy	Supply your own residential proxies to pay the discounted BYO rate. See below.
includePersonalData	boolean	false	Off by default. Product output contains no customer data.

At least one of asins, urls or search is required.

Marketplaces: amazon.com · amazon.co.uk · amazon.de · amazon.fr · amazon.it · amazon.es · amazon.ca · amazon.com.mx · amazon.com.br · amazon.co.jp · amazon.in · amazon.com.au · amazon.nl · amazon.se · amazon.pl

The proxy country is matched to the marketplace automatically.

---

Output reference

One dataset item per product. Every field below is always present; nullable ones are marked.

Field	Type	Null?	Description
asin	string	no	Amazon Standard Identification Number.
url	string	no	Canonical product URL.
title	string	yes	Product title.
brand	string	yes	Brand name.
price	object	no	{ value: number|null, currency: string|null } — current offer.
listPrice	object	no	{ value: number|null, currency: string|null } — RRP / strike-through price.
stars	number	yes	Average rating, 1–5.
reviewsCount	integer	yes	Total ratings count.
inStock	boolean	no	Whether Amazon reports the item in stock.
availability	string	yes	Raw availability text, e.g. "In Stock".
features	string[]	no	Feature bullets. Empty array if none.
images	string[]	no	Absolute image URLs. Empty array if none.
breadCrumbs	string	yes	Category path, e.g. "Video Games > PC > Gaming Mice".
categories	string[]	no	Category path as an array.
bestsellerRanks	object[]	no	{ rank: number, category: string } — Best Sellers Rank per category.
specs	object	no	Amazon's specification table as key/value pairs.
starsBreakdown	object	yes	{ "5star": 79, "4star": 11, ... } rating distribution, in percent.
badges	string[]	no	e.g. "Amazon's Choice", "Prime", "Best Seller".
discountPercent	integer	yes	Computed from price against listPrice.
description	string	yes	Long product description.
boughtInPastMonth	string	yes	Amazon's demand signal, e.g. "10K+ bought in past month".
soldBy	string	yes	Seller name.
shipsFrom	string	yes	Fulfiller, e.g. "Amazon".
deliveryDate	string	yes	Delivery estimate for the zip code you requested.
hasAPlusContent	boolean	no	Whether the listing has A+ brand content.
aspects	object[]	no	Amazon's own review aspects — see below.
deliveryLocation	string	yes	The location the price reflects, e.g. "Beverly Hills 90210".
scrapedAt	string	no	ISO 8601 UTC timestamp.

Example output

{
  "asin": "B07CMS5Q6P",
  "url": "https://www.amazon.com/dp/B07CMS5Q6P",
  "title": "Logitech G305 Lightspeed Wireless Gaming Mouse - Black",
  "brand": "Logitech G",
  "price": { "value": 29.99, "currency": "$" },
  "listPrice": { "value": 49.99, "currency": "$" },
  "stars": 4.6,
  "reviewsCount": 39996,
  "inStock": true,
  "availability": "In Stock",
  "features": ["The next-generation optical HERO sensor delivers incredible performance"],
  "images": ["https://m.media-amazon.com/images/I/31zy93Vq31L._AC_.jpg"],
  "breadCrumbs": "Video Games > PC > Accessories > Gaming Mice",
  "categories": ["Video Games", "PC", "Accessories", "Gaming Mice"],
  "bestsellerRanks": [
    { "rank": 10, "category": "Video Games" },
    { "rank": 1, "category": "PC Gaming Mice" }
  ],
  "specs": { "Brand": "Logitech G", "Color": "Black", "Connectivity Technology": "USB" },
  "starsBreakdown": { "5star": 79, "4star": 11, "3star": 4, "2star": 2, "1star": 4 },
  "badges": ["Amazon's Choice", "Prime"],
  "discountPercent": 40,
  "boughtInPastMonth": "10K+ bought in past month",
  "soldBy": "Sold by Amazon Resale and Fulfilled by Amazon.",
  "shipsFrom": "Amazon",
  "deliveryDate": "FREE delivery Tuesday, September 22",
  "hasAPlusContent": true,
  "aspects": [
    { "label": "quality", "mentions": 3241, "mentionsPercentage": 84, "sentimentMentions": 2715 },
    { "label": "battery life", "mentions": 894, "mentionsPercentage": 84, "sentimentMentions": 754 }
  ],
  "deliveryLocation": "Beverly Hills 90210",
  "scrapedAt": "2026-09-16T19:43:28.633Z"
}

Review aspects

aspects carries Amazon's own AI-extracted review themes with mention counts across the entire review corpus — a sentiment breakdown over tens of thousands of reviews, without scraping a single review body. Useful for product research and competitor comparison.

---

Zip-code accurate pricing

Pass zipCode and the Actor performs Amazon's real delivery-location flow before scraping, so prices, delivery dates and stock reflect that location.

{ "asins": ["B07CMS5Q6P"], "zipCode": "10001" }

Every returned row carries deliveryLocation so you can prove which location a price came from, and deliveryDate — Amazon's delivery estimate for that zip code. Change the zip and both move. Location setup is charged once per run regardless of how many products you scrape.

---

Bring your own proxy

Residential proxy bandwidth is most of what this Actor costs to run, so you can supply your own and pay less.

{ "asins": ["B07CMS5Q6P"],
  "proxyConfiguration": { "proxyUrls": ["http://user:pass@your-gateway:7777"] } }

Residential proxies only — this is not a preference

Amazon blocks datacenter IPs. Measured on 12 live products at production concurrency:

proxy type	requests that returned a product
Residential	83%
Datacenter	25%

A datacenter pool does not fail cleanly — it returns some rows. You would get roughly a quarter of the products you asked for, with the rest silently missing, and the run would look like it worked. Prices and stock you do get are still correct; the problem is everything you never receive.

Worse, a datacenter proxy can pass the startup check by luck, because that check only needs one request to succeed. So the Actor also watches the block rate during the run and warns you if it climbs:

PROXY QUALITY: Amazon blocked 45% of requests through your proxy. That pattern
usually means datacenter IPs — Amazon blocks those (measured ~25% success, against
~83% for residential). Results will be incomplete until you switch to a residential pool.

The same warning is written to RUN_SUMMARY.proxyQualityWarning, so an automated pipeline can detect it without reading logs.

If you are unsure what you have: datacenter proxies are the cheap, fast, unlimited- bandwidth kind. Residential proxies are sold by the gigabyte and cost meaningfully more — roughly what makes this discount worth taking in the first place.

Do not paste an Apify Proxy URL here. proxy.apify.com is not reachable from inside a running Actor, so it fails with unreachable. If you want Apify Proxy, leave this field on "Use Apify Proxy" instead — the Actor wires it up correctly and bills the standard rate.

• Your proxy is checked before the run starts. The Actor fetches an Amazon page through it, retrying on fresh sessions. If it cannot get through, the run stops immediately with the reason and nothing is charged — you are never billed for a run that was doomed by a proxy misconfiguration.
• There is no fallback. If your proxies fail mid-run you get fewer results. The Actor will never quietly finish on our bandwidth at the discounted rate.
• Rotation is yours to provide. One session is pinned to one of your URLs for its lifetime. Supply several URLs, or a gateway that rotates per connection or accepts a session token in the username, or every request will exit from the same IP.
• Leave it unset and the Actor uses our proxies at the standard rate.

Pricing

Event	Price	When
result	$0.002	per product returned, using our proxies
resultByoProxy	$0.0012	per product returned, using your proxies (from 5 October 2026)
locationSetup	$0.02	once per run, only if zipCode is set

The BYO rate takes effect on 5 October 2026 — Apify requires two weeks' notice for any pricing change on a public Actor. You can supply proxies today; until that date those runs are billed at the standard result rate.

• No minimum charge.
• Not-found products are never charged.
• Blocked requests and retries are never charged.

50 products, no location      50 x $0.002            = $0.10
50 products, with zip code    50 x $0.002 + $0.02    = $0.12

---

Use cases

• Price monitoring — track competitor prices per region with zip-code accuracy.
• MAP compliance — verify resellers honour minimum advertised price in each market.
• Dropshipping and arbitrage — check live prices and stock before listing.
• Market research — compare ratings, review distribution, Best Sellers Rank and review aspects across a category.
• Demand estimation — bestsellerRanks plus boughtInPastMonth are the inputs sales-estimate models are built on.
• Logistics and delivery SLAs — deliveryDate per zip code, for comparing fulfilment speed by region.
• Catalogue enrichment — fill in titles, brands, images and features from ASINs.
• AI agents and RAG — feed structured, typed product data to an LLM pipeline.

---

Calling from code

API — start a run and wait for results

curl -X POST "https://api.apify.com/v2/acts/smart_albatross~amazon-product-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"asins":["B07CMS5Q6P"],"zipCode":"10001"}'

JavaScript

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_TOKEN' });
const run = await client.actor('smart_albatross/amazon-product-scraper').call({
  asins: ['B07CMS5Q6P', 'B07L4BM851'],
  zipCode: '10001',
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items[0].price.value);

Python

from apify_client import ApifyClient

client = ApifyClient("YOUR_TOKEN")
run = client.actor("smart_albatross/amazon-product-scraper").call(
    run_input={"search": ["wireless gaming mouse"], "maxItemsPerStartUrl": 20}
)
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["asin"], item["price"]["value"], item["stars"])

---

For AI agents and automated callers

Deterministic contract, in one place:

• Actor ID: smart_albatross/amazon-product-scraper
• Required input: at least one of asins (string[]), urls (string[]), search (string[])
• Mutually exclusive: zipCode and countryCode. Supplying both fails fast, before any network cost.
• Output: one dataset item per product, schema exactly as the Output reference table.
• Never returns: partial rows without an asin, or a numeric 0 standing in for an unknown price.
• Run summary: written to the key-value store under RUN_SUMMARY as { requested, ok, notfound, blocked, charged, locationCharged, blockRate, transport }.
• Failure semantics: a single unavailable ASIN never fails the run. The Actor exits 0 with partial results, and RUN_SUMMARY reports what happened.
• Idempotency: same input, same marketplace and same zip code yield the same rows, except for values Amazon itself changes (price, stock, review counts).

Distinguishing the three "no data" cases:

Case	Row returned?	Signal
Product exists, no current offer	yes	price.value is null, inStock is false
ASIN does not exist / wrong marketplace	no	counted in RUN_SUMMARY.notfound
Amazon blocked the request after retries	no	counted in RUN_SUMMARY.blocked

---

Limits and behaviour

• Search paginates through Amazon's result pages until maxItemsPerStartUrl is reached, the results stop changing, or 20 pages are read — whichever comes first. Amazon serves roughly 30–60 organic results per page, so a request for 500 reads about 10 pages.
• Sponsored placements and availability vary by marketplace and location; results reflect the live page at scrape time.
• price.value is null — never 0 — when Amazon shows no offer.
• Transport retries blocked requests on a fresh session and reports blockRate in RUN_SUMMARY. Retries are not billed.

---

FAQ

Do I need an Amazon API key? No. The Actor reads publicly accessible product pages.

Does it need an Amazon account or login? No. It never logs in, and never asks for credentials or session cookies.

Can I get prices for a specific city or zip code? Yes — that is the point. Pass zipCode and every row reports the deliveryLocation its price came from.

Why is price.value null for some products? Amazon shows no offer for that item — it is unavailable, delisted, or not sold in the selected marketplace. Returning 0 would assert the product is free, so null is used instead. The row still carries title, rating, review count and aspects.

Does it scrape review text? No. It returns review counts and Amazon's aggregated review aspects. It does not return individual review bodies.

Which marketplaces are supported? Fifteen, listed in the Input reference. The proxy country is matched automatically.

What happens if one ASIN fails? Nothing else is affected. The run completes, returns every other product, and records the failure in RUN_SUMMARY.

Why did my run fail with "PROXY CHECK FAILED"? Your proxy could not reach Amazon, so the run stopped before doing any work and you were charged nothing. The message names the cause:

verdict	meaning	fix
auth	the proxy rejected our credentials	check the username and password in the URL
unreachable	no connection at all	check host and port, and that this Actor's IP is allowed by your provider
blocked	reached Amazon, but Amazon blocked it	use residential proxies — Amazon blocks datacenter IPs

---

Legal and compliance

This Actor collects only publicly accessible Amazon product data. It does not log in to Amazon, bypass authentication, or use or request Amazon account credentials or logged-in session cookies. It collects no personal data by default.

You are responsible for ensuring your use complies with applicable law, including GDPR and copyright, and with Amazon's Terms of Service.