Enables robust semantic and full-text search over a directory of markdown files.
Sponsored
This tool functions as an MCP server, offering robust semantic and full-text search capabilities across a designated directory of markdown documents. It intelligently splits markdown files into manageable chunks, generates embeddings using FastEmbed (ONNX, in-process), and stores these vectors along with metadata in LanceDB, which also includes a full-text index. The server ensures your knowledge base is always up-to-date by performing incremental reindexing on startup and automatically watching for file changes.
主要功能
01Real-time file change monitoring and reindexing
020 GitHub stars
03Splits markdown into searchable chunks at headings
04Hybrid semantic and keyword search for markdown content
05Automatic, incremental reindexing of markdown files
06Embeds content using in-process FastEmbed (ONNX)
使用案例
01Integrating contextual knowledge into AI/coding assistants like Cursor or Claude Code
02Building a searchable personal knowledge base from markdown notes
03Enabling efficient information retrieval from project documentation repositories
