Import by Scraping Settings

A scraped page rarely arrives finished. This page, titled Import by scraping in the app, decides what happens to every product the browser extension brings in, so imports come out classified, researched, and written instead of raw. There are four sections. Prepare website imports links to the language and source settings imports rely on. Data model decides whether imports may create attributes and categories. Classification files products into your adopted standards. After an import chains enrichment onto them. Classification runs first, and outside the enrichment chain’s master switch, because the family it assigns decides which attributes enrichment can fill. Where to find it: open your avatar menu > My Settings, then Import & Export > Import by Scraping in the settings sidebar.

​

Prepare website imports

Two links help before your first import:

• Project languages: shows the Primary content language. Set it to the language of the sites you import from, so the content lands in the right language.
• Trusted websites: add manufacturer websites to guide web research.

​

Data model

By default, extension imports fill the attributes and categories you already have. Allow creation only when you want imports to extend this project’s data model. With a switch off, values still go into attributes that exist and breadcrumbs still match category paths that exist. The rest is left out, and the import report says how many spec values and category paths were not imported, and which setting caused it.

​

Classification

If your project has adopted classification standards, imported products are filed into them before anything else runs: The section shows which release you’re on and how many classes or bricks you’ve adopted (for example “ETIM 10.0 · 24 classes”). Until you’ve adopted at least one standard, the toggles stay inert and the section links you to Browse ETIM classes and Browse GPC categories to set that up first.

​

After an import: the enrichment chain

Turn on Enrich scraped products automatically and the steps below run, in order, whenever an extension import finishes.

​

Step 1: Web research

Looks each product up on the web and merges the verified facts first, so the writing step has something real to work from. This costs research credits per product, and enrichment waits for the research to land before writing. This step needs the Pro plan or higher. Without it you can still run web research by hand from the products table. How research behaves (focus, allowed sources, which fields it may update) comes from Web Research settings.

​

Step 2: Write the content

Bulk Enrich writes each field you pick, using the project’s default prompt unless you choose another per field:

• Descriptions
• Short descriptions
• Titles
• SEO title and meta
• Attributes

The first time you switch the chain on, Descriptions, Titles, and SEO title and meta are pre-selected: the three things a scraped page is reliably worst at. Attributes and short descriptions stay off because scrapes often carry usable specs, and every extra field costs credits per product. A summary strip shows the steps that Runs in order, for example web research, then Bulk Enrich with three fields. If nothing is selected, imports stay as they arrive.

​

Related