Shopify Products Scraper & Variant CSV Export

Try two free-to-configure Shopify scraper examples

• Variant SKU and published barcode export — a bounded two-product store sample with CSV/JSON output, SKU, variant barcodes when published, public availability and missing-barcode checks.
• Shopify product price and catalog QA — a two-product catalog example for listed prices, vendors, images and variant counts.

Buyer fit: product feed migration, vendor catalog reconciliation, CSV reporting and scheduled public-listing change checks. This is a permitted public-feed scraper, not a full-store search index; it does not invent currency, barcodes, hidden products or private inventory quantities.

Current usage pricing is shown in Apify's live Pricing tab. Pricing is pay-per-exported-record and may vary by the customer's Apify plan. No login or paid proxy service is required for ordinary permitted public feeds.

Shopify products scraper: CSV/JSON export for catalog operations

Export products and variants from public Shopify products.json feeds for stores you are permitted to inspect. The Actor downloads bounded, robots-permitted public pages and produces product/variant IDs, SKUs, listed prices, visible availability flags, and clean CSV/JSON—without store login, a proxy subscription, or any private inventory access.

Start with a runnable example: Shopify product & variant CSV sample. Review and change its store URL before a live run. Use the published Pricing tab to estimate charges.

Common commercial workflows: Shopify catalog migration, recurring product feed exports, merchandising QA, variant/SKU reconciliation, and price snapshot comparison using stable keys and content_hash. This Actor does not access orders, customers, inventory quantities, unpublished products, protected endpoints or storefronts that forbid the public feed.

If you need another operation: Shopify cross-store product search helps filter published feeds by SKU, brand and title; Shopify product/variant change monitor compares snapshots. These are separate Actors with separate pricing.

Export the public /products.json feed of a Shopify store (or one collection) as clean product and variant rows, in the dataset (JSON) and as spreadsheet-safe CSV files. It uses no login, no proxy and no paid API.

New: merchandising insights and store diagnostics

Each successful export now writes a CATALOG_INSIGHTS JSON summary in the Key-value store and embeds it in SUMMARY. For each store you can see:

• Distinct exported product and variant counts (both and variant-only output modes supported).
• Share of available/unavailable variants, missing SKUs, and published discount signals.
• Most frequently exported vendors and min/max listed prices per store.
• Error/blocked-store explanations with safe follow-up steps, without bypassing robots.txt.

This enables recurring Shopify merchandising audits and catalog migration QA without requiring an external spreadsheet just to count missing SKUs or summarize the catalog. The report uses the same rows already exported, so it causes no additional storefront HTTP requests and no new billable events. Results reflect only the bounded public feed and selected output mode; if you select product rows only, SKU-level metrics are unavailable. Product currency and private inventory quantities are not inferred.

What makes the output useful

• Stable identifiers: product_key (<host>/<product_id>) and variant_key (<host>/<variant_id>) stay the same between runs, so you can join exports from different days.
• Change detection: content_hash is a short hash of tracked fields (title, SKU, options, price, compare-at price, availability, weight). Compare it between runs to find changed rows.
• Normalised prices: decimal strings with two places, plus a computed discount_percent when a compare-at price is higher.
• Variant-level detail: option names and values, SKU, availability, weight, shipping/tax flags, variant image and a direct ?variant= URL.
• Spreadsheet-safe CSV: text starting with =, +, - or @ is prefixed with ' so it cannot run as a formula. The file starts with a UTF-8 BOM for Excel.
• Bounded and polite: hard caps on products and pages, a delay before every request, and a stop if the feed repeats itself.

Safety and compliance behaviour

• Only https:// store URLs with public DNS names are accepted. IP addresses, local names, credentials in URLs and non-default ports are rejected, and resolved addresses must be public.
• robots.txt is read first and every feed URL is checked against it (wildcards, $, longest match, Crawl-delay up to 30 s). If robots.txt cannot be read (anything other than 200, 404 or 410), nothing is requested.
• Redirects are followed only to the same host and path. Responses are size-limited.
• Only the public feed is read. No customer, order or personal data is collected.

You are responsible for having the right to read and use the data of each store you export, and for following its terms.

Sample input

{
    "storeUrls": ["https://www.allbirds.com"],
    "exportMode": "variants",
    "maxProductsPerStore": 100,
    "maxPagesPerStore": 5,
    "pageSize": 250,
    "onlyAvailableVariants": false,
    "includeDescription": false,
    "requestDelayMs": 1000
}

The store above is only a format example, not a statement about its policies. A collection works too, for example https://www.allbirds.com/collections/mens. Use the store's final public domain.

Field	Default	Meaning
storeUrls	required	1 to 5 store or collection URLs
exportMode	variants	variants, products or both
maxProductsPerStore	100	1 to 10000
maxPagesPerStore	5	1 to 100
pageSize	250	1 to 250
onlyAvailableVariants	false	Skip variants not reported as available
includeDescription	false	Add plain-text description to product rows
descriptionMaxChars	500	0 to 5000
requestDelayMs	1000	500 to 10000

Output

• Dataset: one record per variant and/or product (record_type). Views for variants and products are defined in the dataset schema.
• Key-value store: variants.csv and/or products.csv (matching the export mode) and SUMMARY (per-store status, pages, counts, stop reason, truncation flag).

Variant columns include: variant_key, product_key, sku, variant_title, option1_name/option1_value (up to 3), price, compare_at_price, discount_percent, available, grams, requires_shipping, taxable, image_url, variant_url, content_hash.

Product columns include: product_key, title, vendor, product_type, tags, option_names, variant_count, available_variant_count, price_min, price_max, image_count, product_url, published_at, updated_at, content_hash.

Store status values

complete (reached the end of the feed or your caps), partial (stopped on an error after some data), blocked (unsafe host or robots.txt), failed. truncated: true means more products probably exist beyond your caps.

Billing behaviour in the code

The code emits one event named product-exported per product from stores whose status is complete. Partial, blocked and failed stores are not counted. This repository does not configure a payable plan or a price; any pricing must be set up separately in the platform and checked against real run cost first.

Limitations

• Many stores disable or restrict /products.json, use a different domain, or rate-limit; those runs end as blocked or failed with a reason.
• The public feed has no currency, no stock quantity, no cost data and no unpublished or draft products. Prices are the displayed storefront values and may differ by market or currency.
• The feed order and content can change while paging, so products may be missed or repeated in a very active catalog. Repeats are dropped by product ID.
• DNS is checked before requests, but the HTTP library resolves the name again when connecting, so DNS rebinding is reduced rather than fully eliminated.
• A single feed page can be up to 12 MB; larger responses are rejected.
• A 429 or 503 response is retried once, then the store ends as partial or failed.
• Market-size figures for similar tools (about 105 monthly active users, free users included) say nothing about paid demand. This Actor has not been validated with paying customers, and no sample results are published here.

Barcode and GTIN coverage for product feed exports

Variant CSV and Dataset records now include the Shopify barcode field when the public products.json feed provides one. It may contain a UPC, EAN or another store-defined value; the Actor does not invent GTINs or validate barcodes. The automatic CATALOG_INSIGHTS report also counts variant rows with missing barcodes, helping catalog migration teams identify fields that need review.

Existing variants without a published barcode remain blank. This feature uses the same requests and event charges as the normal export.