Skip to main content
Read one storefront product page and receive the product as JSON through the async task API. Pass the page’s url; the URL picks the storefront, and every storefront returns the same fields. A completed read costs 3 credits.

​
Request

Submit the task, then poll GET /v1/async/task/{id} or receive its webhook. See Async tasks.

​
Parameters

  • url: Required. A product page on one of the storefronts below, up to 2,048 characters. Product url values that Commerce Search returns for these storefronts work as they are, except Yahoo! Shopping’s LOHACO pages on lohaco.yahoo.co.jp. Parameters the result does not depend on, such as tracking codes, are dropped before the page is read.
  • country: Optional. Defaults to the storefront’s country. Any other country is rejected with 422 REGION_UNSUPPORTED.
Other addresses of the same page also work, such as Amazon’s /gp/product/{ASIN} and 11st’s mobile pages.

​
Response example

​
Result fields

  • site, country, url and productId identify the page. url is the page that was read. productId is the storefront’s own product ID, such as the ASIN on Amazon or shop:item on Rakuten, and matches the one Commerce Search returns. The exception is a Nykaa product with shades or sizes: Commerce Search returns the ID of one option, and this read returns the ID of the product page.
  • title is always present. The other fields appear when the page shows them; the table below lists the storefronts that return each one.
  • price is what a shopper pays for the product as the page first shows it, before coupons and delivery, in currency (ISO 4217). listPrice is the regular price the page strikes through, present when it is above price. inStock is that product’s stock status, and seller is the merchant selling it.
  • rating uses a five-point scale with two decimals. reviewCount is the number of reviews.
  • images lists up to 30 photos from the product’s gallery, in page order. 11st and Lotte ON return the main photo only, Sephora up to two, and on Rakuten and Yahoo! Shopping the banners or notices a seller places in that gallery come along. description is plain text with one line per paragraph, up to 10,000 characters; text that appears only inside images is not included.
  • descriptionImages lists up to 300 images from the product description, such as a seller’s detail page or Amazon’s A+ content, in page order. Banners and notices that sellers place in their description come along, and text inside the images is not read.
  • specs lists up to 100 { name, value } pairs from the product details. options lists up to 20 choices such as size or color, each with up to 100 values; the other fields describe the choice the page opens on. categories runs from the broadest category to the narrowest, up to 10 levels. gtin is the UPC, EAN or JAN the page lists.
  • google_shopping reads a product that Commerce Search returned: its url carries the search query, Google’s product ID and the product’s title. The read runs that query on Google Shopping again, or the title when Google no longer lists the product for the query, and opens the product’s panel, so it takes longer than a storefront page. offers lists the stores Google shows for the product, starting with the one Google marks as the best price, each with seller, title (the store’s product name), url (the store’s product page), price, listPrice, currency and inStock. The product’s price, listPrice, seller and inStock are the first store’s. reviewCount appears only when Google shows the exact count (below 1,000). If Google lists the product for neither, the task ends with PAGE_UNAVAILABLE.
  • A page the storefront answers with 404 or 410, such as a removed product, ends the task with PAGE_UNAVAILABLE, and the message carries the storefront’s answer. Other failed reads end with ENGINE_NO_RESULT. Billing applies per completed read.