Carrot Search

Carrot2 Search (search.carrot2.org) is an open-source search results clustering engine that takes results from standard search engines (Google, Bing, others) and automatically groups them into meaningful topic clusters, revealing the thematic structure of a set of search results through AI-powered categorization.

For OSINT investigators, Carrot2's clustering approach provides a different perspective on search results than a standard ranked list. When a name, term, or concept returns many search results covering multiple unrelated topics, Carrot2 automatically groups results into clusters — helping investigators quickly identify which cluster of results is most relevant to their investigation while understanding the full landscape of what the term covers.

This clustering is particularly valuable when investigating common names or ambiguous terms that generate large, heterogeneous result sets. A search for a common name might return results about a politician, an athlete, a business executive, and a fictional character. Carrot2 clusters these automatically, allowing investigators to immediately focus on the relevant individual's results.

For topic research at the beginning of a new investigation, Carrot2's cluster visualization provides a rapid orientation to what aspects of a topic are covered by public web sources — similar to Answer The Public but for actual search results rather than search queries.

The tool supports multiple clustering algorithms with different trade-offs between cluster granularity and coherence, allowing investigators to adjust the clustering approach for different types of content.

Carrot2 is also used in academic research for search result analysis and in competitive intelligence for understanding the structure of competitor-related web content.

Use Carrot2 as an orientation tool at the beginning of broad investigations to identify which clusters of results are most relevant before conducting targeted deep searches on specific topics.