01High-accuracy Japanese full-text search using SudachiPy for precise tokenization.
02Indexes local text files, supporting recursive directory scanning, `.gitignore` filtering, and extension-based targeting.
03Automates index cleanup by removing entries for deleted files during re-indexing.
04Provides precise line numbering for search results within tokenized Japanese text using a token map strategy.
050 GitHub stars
06Integrates with FastMCP to expose directory indexing and document search as accessible MCP tools.