Ultimate Web ScraperScrape Product Data with Zero Configuration
Automatic Extract reads a page's structured data — JSON-LD, microdata, meta tags — and returns names, prices, SKUs, and ratings with no selectors.
Most modern e-commerce sites (and news sites, and job boards) embed structured data in every page for search engines. The Page Extractor's Automatic Extract step reads it directly — no element picking, no selectors, no configuration.
Collect the product URLs
You need a list of product page URLs. The fastest source is usually the Sitemap Explorer: open it on the store, let it discover the sitemap, and select the products branch of the URL tree. Alternatively, scrape a category page with the List Extractor or upload a CSV.
Load them into the Page Extractor
Open Page Extractor and add your URLs in Step 1 — via Sitemap Explorer, Data Source, or CSV upload. The first URL loads in the active tab.
Let Automatic Extract take over
If the first page contains structured data, the tool detects it and auto-adds the Automatic Extract step — you'll see it appear in Step 2 without doing anything. It even probes a few times to catch structured data that's injected late by JavaScript. If it doesn't appear automatically, you can add it manually.
Run and review
Start the extraction. Each page becomes one row with columns drawn from the page's structured data. For products, that typically includes:
| Column group | Examples |
|---|---|
| Identity | Name, Brand, Category, Type |
| Pricing | Price, All Prices, Currency, Offers count |
| Stock | Availability |
| Attributes | Color, Size, Material, Weight |
| Social proof | Rating (e.g. "4.6 (1,203)"), Reviews |
| Identifiers | SKU, GTIN-8/12/13/14, MPN, Variant SKUs |
| Media | Image(s) |
Prices and enumerations are normalized, and cross-referenced entities are resolved for you.
How variants are handled
Product variants consolidate into one row per URL — sizes, colors, and per-variant prices are joined into multi-value cells rather than exploded into separate rows. The complete raw structured data is preserved in a _raw_jsonld column if you need every detail.
Where it works — and where it doesn't
Automatic Extract is generic: it reads JSON-LD, microdata, and Open Graph/Twitter/standard meta tags, so it works on WooCommerce, Magento, custom stores, news sites, job boards — anywhere structured data exists. It picks the page's subject entity (Product, Article, JobPosting, Event, Recipe, LocalBusiness, and more) and flattens it into a row.
Its limits are the flip side of the same design:
- Sites without structured data return little or nothing. Use the element picker in the Page Extractor to define fields manually instead — both steps can coexist in one run.
- It reads markup, not JavaScript state: data that lives only in JS variables isn't captured.
- Listing pages (a page describing many products) collapse to one row — feed it detail pages, not category pages.
🛍️ Scraping a Shopify store?
For a whole Shopify catalogue, skip this workflow entirely — the dedicated Shopify Extractor pulls the full catalogue in one pass, one row per variant. See Export any Shopify store to CSV.