Provides Retrieval Augmented Generation (RAG) capabilities for your markdown documentation.
Sponsored
Inkdex acts as a powerful RAG system designed specifically for your markdown documents, making your knowledge base semantically searchable. It operates by parsing markdown files, segmenting them into meaningful chunks based on heading structure, and then generating local embeddings using a robust model like `all-MiniLM-L6-v2`. These embedded chunks are efficiently stored in SQLite. For search, Inkdex employs a sophisticated hybrid approach, combining vector similarity with BM25 full-text matching, and intelligently fuses the results using Reciprocal Rank Fusion to deliver highly relevant search outcomes. The tool is seamlessly exposed over the Model Context Protocol (MCP), allowing easy integration into various MCP clients for enhanced documentation interaction.
Key Features
01Generates local embeddings using `all-MiniLM-L6-v2`
