Diffbot

Diffbot uses machine learning and computer vision to read web pages the way a person would. Its extraction APIs identify the page type, such as an article, product or discussion thread, and return the content as structured fields without site-specific rules.

Its Analyze API decides the page type for you, while Article, Product and other APIs return fields like title, author, date, price and images. Crawlbot applies extraction across a whole site. Diffbot also builds a large Knowledge Graph of organizations, people and articles drawn from the public web.

It suits teams that collect from many different sites and cannot write a parser for each one. Check extraction quality on a sample of your own target pages, since unusual layouts can still produce missing or wrong fields.