Fast regex Matching with Indexing

Cursor details how agentic coding has revived regex search as a core context-gathering primitive, and explains their work building a faster regex-over-codebase search layer on top of ripgrep using indexing techniques. The post explores why semantic search alone is insufficient for agents — certain queries require exact regex matching — and how indexing can dramatically accelerate what would otherwise be a full-scan operation. Directly relevant to AI coding agent infrastructure and context retrieval strategies.