web-extract

Structured data extraction from web pages without parsing raw HTML is done via a single API call that returns ready-to-consume JSON fields like title, summary, sections, key metrics, and outgoing links. This eliminates the need for a second LLM pass to parse markdown or raw HTML.

Yes, you can scrape JavaScript-rendered pages and interact with SPAs. The extraction process supports interactive actions via click, scroll, and write operations to fully render and capture dynamic content before returning the structured JSON output.

The best way to crawl an entire website with path depth filters is using asynchronous recursive crawling. This method discovers domain URLs and systematically extracts structured JSON data across the site while applying depth and path filters to control the scope of the crawl.

Yes, you need a valid ZOODATA_API_KEY credential to perform Google SERP searches with deep scraping. This key is required to authenticate all API calls, including single URL scraping, domain URL discovery, and asynchronous recursive site crawling.

Web scraping with this method returns article metadata including author, date, and page type by default. The structured JSON output automatically extracts these specific fields alongside the page title, summary, key sections, and outgoing links.

Structured data extraction requiring a second LLM pass happens when using basic scrapers that return raw markdown or HTML. Using a dedicated extraction API returns default structured JSON fields directly, eliminating the need for post-scrape parsing and extra LLM passes.