Costco Scraper

Extract costco.com product data for price monitoring, MAP enforcement and warehouse-vs-online arbitrage research: product names, item numbers, online and warehouse prices, Instant Savings promos, stock status, ratings, review counts, images, specs and category paths. Keyword search runs against Costco's own internal search API at $0.85 per 1,000 products with no browser and no proxy, and optional customer reviews from Costco's Bazaarvoice feed cost $1.00 per 1,000 products.

Main Use Cases

• Costco price monitoring - track online and per-warehouse prices plus Instant Savings promos over time.
• MAP and dealer price checks - compare costco.com online pricing against warehouse pricing for a list of items.
• Warehouse vs online arbitrage research - find items that are cheaper in one channel.
• Product catalog and SKU data builds - item numbers, catalog IDs, brands, specs, images and category paths.
• Customer review mining - pull reviews, star ratings, verified-purchase badges and helpful-vote counts.
• Stock and availability tracking - monitor InStock / OutOfStock per warehouse and per ship-to ZIP.

How It Works - The Bypass (No Official Costco API, No Browser)

Costco's storefront is a Next.js app; its listings and per-warehouse prices load client-side from internal JSON APIs that sit behind Kasada Bot Defense. The Actor reverse-engineers the whole data surface and uses two of its endpoints:

• Keyword search calls Costco's internal GRS search API directly. No browser and no Kasada token are required when the request carries the storefront's public client-identifier header and a complete JSON body with a non-empty visitorId.
• URL mode parses server-rendered product pages, which embed application/ld+json Product + BreadcrumbList plus a priceInfo state blob (online/delivered price, discounts, promo).

#	Surface probed	Result
1	Internal search API POST https://gdx-api.costco.com/catalog/search/api/v1/search	USED (keyword mode): returns real product rows without any Kasada token, provided (a) the request carries the public client-identifier header and (b) the JSON body is complete and includes a non-empty visitorId. Missing header returns 401 Missing credentials; missing visitorId or fields returns 400 Bad Request; a partial body returns the Kasada error page. With the full body it is a plain 200.
2	Embedded JSON in served PDP HTML	USED (URL mode): every /p/-/<slug>/<id> page is server-rendered and embeds application/ld+json Product + BreadcrumbList plus a priceInfo state blob.
3	Mobile gateway https://gdx-api-mobileapps.costco.com/.../search	Same API and behaviour as #1 (same GRS backend); not needed now that the web host works directly.
4	Product API .../catalog/product/product-api/{v1,v2}/products	403 with the search credential, 401 Missing credentials without it; still gated.
5	Warehouse shelf price dispprice-api	Superseded: in search mode the per-warehouse price and stock already come back as inventory(<warehouseId>, price) and availability attributes inside the search response, with no separate call.
6	Kasada p.js VM (/149e9513-.../2d206a39-.../p.js, 152 KB, zvm-obfuscated)	Not needed. Kept for reference: it is bound to IP + User-Agent + TLS fingerprint, so forging tokens server-side is not viable, but the search API does not require them.
7	Product GraphQL https://ecom-api.costco.com/ebusiness/product/v1/products/graphql	Needs member OAuth (costco-x-authorization: Bearer ...).
8	Third-party search backend (Algolia / Meilisearch / ...)	None: Costco uses its own GRS / Fusion stack.
9	Bazaarvoice reviews https://apps.bazaarvoice.com/bfd/v1/clients/Costco/.../reviews.json	USED (reviews): plain JSON with a public bv-bfd-token, paginated (limit/offset, cap 100 per page), returning about 50 fields including text, rating, helpful votes, verified-purchase badge and location.

Why this choice: the internal search API returns 12 of 12 fields as clean JSON with zero anti-bot cost when called the way the storefront calls it. No browser, no proxy, no paid solver. The client-identifier is a public value embedded in the costco.com bundle and in the Actor's code as DEFAULT_CLIENT_IDENTIFIER (not a secret and not an input field). If Costco rotates it, update the constant in the Actor source; grep the served HTML/JS for client-identifier.

Input Parameters

Field	Type	Required	Default	Notes
searchQueries	array (string)	No	["monitor"] prefill	Keywords, one per line. When set, the Actor runs search mode against the internal GRS search API and returns matching products. Leave empty to scrape explicit product URLs instead. Results are limited by maxItems.
productUrls	array (string)	No	one monitor PDP prefill	Direct Costco product page URLs, e.g. https://www.costco.com/p/-/..., one per line. Used only in URL mode.
warehouseId	string	No	847	Costco warehouse number used for local price and stock.
shipToPostal	string	No	98101	Delivery ZIP code used to resolve online availability and pricing.
shipToState	string	No	WA	Two-letter state code matching the ZIP.
includeReviews	boolean	No	false	Fetch customer reviews for each product from Costco's Bazaarvoice feed into a separate reviews dataset. Enabling this applies an extra pay-per-event charge (reviews) once per product.
maxReviewsPerProduct	integer	No	50 (1-1000)	Maximum number of customer reviews to pull per product while includeReviews is enabled.
maxItems	integer	No	100 (1-20000)	Maximum number of products saved to the dataset; also sets the search page size.
maxRequestsPerCrawl	integer	No	300 (min 1)	HTTP safety cap when crawling explicit product URLs. Not used in search mode.
proxy	object	No	useApifyProxy: false	Standard Apify proxy picker, off by default. Search mode does not need a proxy.

Free Apify users get a 10-item preview per run; paid users get the full maxItems budget.

Output Fields

Each product row (default dataset) carries recordType: "product". Fields:

• Identity and classification: recordType, source (search or crawl), url, productName, itemNumber, catalogId, brand, model, category, categoryPath (breadcrumb array), description, specs (array).
• Pricing: listPrice (pre-discount), onlinePrice (effective post-discount price), deliveredPrice (when reported separately), warehousePrice (selected warehouse, search mode only, null in URL mode), discountAmount (listPrice - onlinePrice, 0 when no promo), currency.
• Promotions: promo (text when an Instant Savings offer is active), promoStart, promoEnd, promoDetails (raw blob for analysts).
• Availability and ratings: availability (e.g. InStock, OutOfStock), rating (0-5), reviewCount.
• Media: images (gallery array), primaryImage.
• Run context: searchQuery, warehouseId, scrapedAt.

Reviews dataset (optional, separate)

When includeReviews is on, each product's reviews go to a separate, run-scoped reviews dataset (declared under storages.datasets in actor.json, each dataset with its own schema). One row per review, with recordType: "review" and source: "bazaarvoice":

itemNumber, productName, reviewId, title, reviewText, rating, author, authorLocation, submissionDate, isRecommended, isVerifiedPurchase, positiveFeedbackCount, negativeFeedbackCount, totalFeedbackCount, language, pros, cons, scrapedAt

On the run page, open the Storage tab and pick the dataset from the Dataset dropdown to view or export each one. The run Result tab links both (Products and Reviews). If opening the reviews dataset ever fails, review rows are written to the default dataset with recordType: "review" instead and a warning is logged.

Billing: only product rows in the default dataset trigger the built-in apify-default-dataset-item event. Review rows in the non-default dataset are charged through the custom reviews pay-per-event (1 event per product).

Example Input And Output

Example input (keyword search with reviews):

{
  "searchQueries": ["monitor"],
  "maxItems": 50,
  "warehouseId": "847",
  "shipToPostal": "98101",
  "shipToState": "WA",
  "includeReviews": true,
  "maxReviewsPerProduct": 30
}

Example product row:

{
  "recordType": "product",
  "source": "search",
  "searchQuery": "monitor",
  "url": "https://www.costco.com/p/-/msi-27-gaming-qhd-qd-oled-240hz-003ms-monitor/4201021105",
  "productName": "MSI 27\" Gaming QHD QD-OLED 240Hz 0.03ms NVIDIA G-Sync & AMD FreeSync Monitor",
  "itemNumber": "2056897",
  "catalogId": "4201021105",
  "brand": "MSI",
  "model": "MAG 271QPX",
  "listPrice": 399.99,
  "onlinePrice": 349.99,
  "warehousePrice": 339.99,
  "deliveredPrice": 349.99,
  "discountAmount": 50.0,
  "currency": "USD",
  "promo": "Instant Savings $50 off",
  "availability": "InStock",
  "rating": 4.88,
  "reviewCount": 8,
  "primaryImage": "https://mobilecontent.costco.com/live/resource/img/...",
  "categoryPath": ["Computers", "PC Gaming", "Gaming Monitors"],
  "category": "Gaming Monitors",
  "warehouseId": "847",
  "scrapedAt": "2026-10-11T09:30:00.000Z"
}

Example review row (reviews dataset):

{
  "recordType": "review",
  "source": "bazaarvoice",
  "itemNumber": "4000287902",
  "productName": "LG 32\" Class Ultrafine 4K UHD IPS Monitor",
  "reviewId": "4C5530F5-12345",
  "title": "Great monitor!",
  "reviewText": "Crisp text and excellent color out of the box.",
  "rating": 5,
  "author": "Scott",
  "authorLocation": "Seattle, WA",
  "submissionDate": "2026-09-01",
  "isRecommended": true,
  "isVerifiedPurchase": true,
  "positiveFeedbackCount": 4,
  "negativeFeedbackCount": 0,
  "totalFeedbackCount": 4,
  "language": "en_US",
  "pros": ["Sharp text", "Good color"],
  "cons": ["No speakers"],
  "scrapedAt": "2026-10-11T09:31:00.000Z"
}

Integrations & Automation

• Apify API - run via POST https://api.apify.com/v2/acts/{actorId}/runs, export via GET /v2/datasets/{datasetId}/items?clean=true&format=json.
• Schedules - run daily to build a costco.com price history (recurring usage also improves Store recommendation signals).
• Webhooks - fire on run finished to push new price drops into your own system.
• Zapier / Make - via the Apify integration, e.g. new Instant Savings offers into Google Sheets or Airtable.
• Pay-per-event - $0.85 per 1,000 products (result), plus $1.00 per 1,000 products when includeReviews is on (reviews, charged once per product), plus a $0.00005 run start.

Related Actors

• costco-fast-product-scraper: the Apify-owned Costco product scraper, useful for a like-for-like feature comparison.
• costco-reviews-scraper: Apify-owned Costco reviews actor, if you need reviews only.
• rigelbytes/costco-scraper: a smaller community Costco product actor.
• tesco-scraper: another large grocery / retail product scraper with price and detail fields.
• amazon-bestsellers-scraper: Amazon category bestseller rankings for cross-retail product research.

FAQ

Does Costco have an official API?

Costco has no public product API for third parties. The member-facing GraphQL (ecom-api.costco.com) needs member OAuth, and the product API endpoints return 403. This Actor uses Costco's own internal GRS search API the same way the storefront does, plus server-rendered PDP JSON in URL mode, so no key is required.

What are the alternatives to this Costco scraper?

Store alternatives include e-commerce/costco-fast-product-scraper (about $0.0043 per product detail page), rigelbytes/costco-scraper (about $0.001 per result), parseforge/costco-scraper ($0.00757 per result) and e-commerce/costco-reviews-scraper ($0.002 per item, reviews only). This actor is the cheapest product mode seen at $0.85 per 1,000 results and adds per-warehouse pricing and Instant Savings fields.

How do I get Costco warehouse prices?

Set warehouseId (default 847). In search mode the internal API returns the per-warehouse price and stock inside the same search response, so warehousePrice is populated. In URL mode there is no warehouse lookup, so warehousePrice is null.

Do I need a proxy or a browser to scrape Costco?

No. Search mode is plain HTTPS with no browser and no proxy. URL mode is Cheerio-only. Kasada tokens are not required for the search API when the request is formed the way the storefront forms it.

How do I get Costco customer reviews?

Set includeReviews: true. Reviews are fetched from Costco's Bazaarvoice feed into a separate reviews dataset, up to maxReviewsPerProduct per product. This adds a reviews pay-per-event charge once per product.

How much does a Costco scrape cost?

$0.85 per 1,000 product rows in the default dataset ($0.00085 per row), plus $1.00 per 1,000 products when includeReviews is on (one reviews event per product), plus a $0.00005 run start. Runs fit comfortably in 256 MB and need no paid anti-bot service.

Can I track Costco prices on a schedule?

Yes. Create an Apify Schedule for a daily run with your searchQueries and warehouseId, and store each day's dataset to build a price history and to catch Instant Savings changes.

Why is warehousePrice null sometimes?

warehousePrice is only returned in search mode, because only the internal search API carries per-warehouse inventory. In URL mode (productUrls) that field is null by design.

SEO Keywords

costco scraper, costco api, costco product data, costco price tracker, costco warehouse prices, costco reviews scraper, costco item number lookup, costco inventory checker, costco price monitoring, costco product catalog, costco instant savings, costco price comparison, warehouse club pricing data, costco map enforcement, costco sku data, costco stock checker, scrape costco, costco dataset, costco price history, costco product images, costco online vs warehouse price

For AI Agents & LLM Apps

• Purpose: returns one JSON row per costco.com product (name, item number, catalog ID, brand, online / warehouse / delivered prices, discounts, promo, availability, rating, review count, images, category path) and, optionally, one row per customer review in a separate dataset.
• Minimal working input:

{ "searchQueries": ["monitor"], "maxItems": 10 }

• Variant input (reviews enabled, ship-to a specific ZIP and warehouse):

{ "searchQueries": ["monitor"], "warehouseId": "847", "shipToPostal": "98101", "shipToState": "WA", "includeReviews": true, "maxReviewsPerProduct": 30 }

• Output field names: recordType, source, searchQuery, url, productName, itemNumber, catalogId, brand, model, listPrice, onlinePrice, deliveredPrice, warehousePrice, discountAmount, currency, promo, promoStart, promoEnd, promoDetails, availability, rating, reviewCount, images, primaryImage, categoryPath, category, description, specs, warehouseId, scrapedAt.
• Review row field names: recordType, source, itemNumber, productName, reviewId, title, reviewText, rating, author, authorLocation, submissionDate, isRecommended, isVerifiedPurchase, positiveFeedbackCount, negativeFeedbackCount, totalFeedbackCount, language, pros, cons, scrapedAt.
• Behaviors an agent should know:
  • Mode priority: if searchQueries is non-empty the Actor runs search mode; productUrls is used only when there are no search queries. Do not set both expecting a merged result.
  • warehousePrice is populated only in search mode and is null in URL mode.
  • onlinePrice is the effective post-discount price; listPrice is pre-discount; discountAmount = listPrice - onlinePrice (0 when no promo).
  • includeReviews: true adds a billed reviews event once per product and writes to a separate reviews dataset (not the default dataset); if that dataset cannot be opened, review rows fall back to the default dataset with recordType: "review".
  • maxItems also sets the search page size, so a very small value can reduce how many rows fit per request.
  • maxRequestsPerCrawl is ignored in search mode.
• Billing: pay-per-event - $0.00085 per product row in the default dataset, $0.001 per product when reviews is enabled (one event per product), $0.00005 per run start. Tiered pricing may be lower on higher Apify plans.

Legal & Compliance Disclaimer

This actor is an independent tool and is not affiliated with, endorsed by, or sponsored by Costco Wholesale Corporation (costco.com). It accesses only publicly available product pages and Costco's own storefront search API, does not log in, does not bypass member authentication and does not solve CAPTCHAs.

Users are responsible for complying with costco.com's terms of service and applicable data-protection law. Review rows can include a reviewer nickname and an optional review location; do not use extracted personal data for unsolicited commercial outreach in violation of applicable law (for example CAN-SPAM, GDPR or CCPA). Use the data for lawful purposes such as price research, catalog building and market analysis.