Amazon Search Scraper - Structured Results

Overview

The Amazon Search Scraper - Structured Results preset is a fully automated Figranium browser task designed to query Amazon and extract detailed item lists from the first results page. Running in agent mode with built-in stealth behaviors, rotate user agents, and human typing simulation, it navigates directly to target product queries on www.amazon.com without triggering bot protection mechanisms.

What it does

• Navigates to Amazon using a parameterized search URL configured via the query variable.
• Simulates human typing and randomized cursor movements while waiting for search result cards (div[data-component-type="s-search-result"]) to render.
• Executes a controlled scroll action down the page to trigger lazy-loaded images and dynamic DOM elements.
• Extracts a structured list of items via custom JavaScript, filtering out empty cards or non-product slots.
• Returns precise product attributes including rank, ASIN, title, pricing, ratings, reviews, absolute URLs, image sources, availability, and sponsored labels.

Inputs

• query: A text variable specifying the product category or keywords to search for on Amazon (defaults to wireless headphones).

Output

• Returns a clean JSON dataset containing structured product records with the following fields:
  • rank: The numerical position of the product on the search page.
  • asin: The unique Amazon Standard Identification Number.
  • title: The full product title.
  • price: The normalized product price.
  • currency: The currency code (set to USD).
  • rating: The extracted numerical product rating.
  • reviews: The total review count.
  • url: The direct product page URL.
  • image_url: The primary product thumbnail image source.
  • availability: Availability status text when present.
  • sponsored: A boolean flag indicating whether the item is a sponsored placement.

How to use

1. Open the preset in your Figranium workspace.
2. Locate the query variable in the task configuration panel and update its value to your desired search term.
3. Run the task. The agent will launch a secure browser instance, perform the search, scroll to load all elements, and extract the structured data automatically.
4. Export or consume the JSON extraction results directly from the run summary.

Notes

• This task uses advanced stealth settings including overscroll, cursor glide, and natural typing to ensure high reliability.
• Execution is stateless and designed to pull data exclusively from the initial search results page matching your query.