01Hybrid Code Search combining BM25 and dense vector search for comprehensive results.
02Efficient incremental indexing of only changed files using Merkle trees.
03Fixed snapshot persistence and improved background sync for reliable data.
04Context-aware indexing that understands code relationships across millions of lines.
05Automatic Milvus flush ensures immediate data availability after indexing.
061 GitHub stars