Codemogger

Code indexing library with tree-sitter parsing and semantic search via embedding-based similarity.

Indexes codebases using tree-sitter for structural parsing across 13 programming languages, combined with semantic search powered by all-MiniLM-L6-v2 embeddings stored in SQLite. Extracts functions, classes, and type definitions with full context, enabling natural language queries against code structure and meaning. Designed as a library that can be embedded in development tools and AI coding assistants.