Securely indexing large codebases

This post details the implementation of secure codebase index sharing using Merkle trees for change detection and simhash-based vector search for reusing existing indexes within a team. It introduces content proofs derived from Merkle trees to guarantee that clients only access files they possess, ensuring security and privacy. The system significantly reduces time-to-first-query for large codebases by enabling immediate semantic search against a copied index while background synchronization reconciles differences.