Taxonomies are well accepted tools for large-scale text indexing. But taxonomies in practice are implemented as shallow hierarchies for navigation and categorization, or as controlled dictionaries and thesauri – all using only a string as the terminal concept representation. This paper presents an approach called orthogonal corpus indexing (OCI) which generates large scale concept indexes from existing text corpora. A concept is a richly multifaceted structure including a title, a signature vector of facet-weights, and links within in a parent-child topic hierarchy. Based on content from reference publishers and public domain sources, Intellisophic is actively building an extensible library of taxonomic content. Currently, this library covers a broad set of subject areas with millions of contextually precise and well articulated concepts. This paper outlines the company’s content, methods, and systems.