Diffbot

Diffbot is a web data extraction and knowledge graph company founded in 2008 by Mike Tung, a Stanford AI Lab researcher. The company uses computer vision and natural language processing to automatically understand and extract structured data from web pages.

Diffbot's core technology differs from traditional web scrapers by using AI to visually parse web pages the way a human would — identifying articles, products, discussion threads, and other content types without requiring custom rules for each website. The company maintains the world's largest commercially available knowledge graph, containing over 20 billion entities and 1 trillion facts extracted from the entire crawlable web, updated continuously.

Diffbot's Knowledge Graph API and web extraction APIs serve AI companies, search engines, and enterprises that need structured data at scale. In the AI era, Diffbot's technology is particularly valuable for building RAG systems, training datasets, and real-time web intelligence — providing the kind of structured, up-to-date world knowledge that makes AI applications more accurate and useful.