Local RAG
Local RAG for semantic document search without external API calls.
A document search server that runs entirely on your machine. Supports PDF, DOCX, TXT, and Markdown files. Uses Transformers.js for generating embeddings locally and LanceDB for vector storage. No external API calls after initial model download, no cloud services, works offline.
Provides semantic search capabilities through natural language queries. Returns results in under 3 seconds for typical document collections. Designed for developers who need document search functionality while keeping data on their local machine. Setup via npx with minimal configuration required.
Source
Repository: https://github.com/shinpr/mcp-local-rag
Maintain Local RAG?
Let people know it's listed here — add the badge (live metrics, light/dark aware) or a plain link to your README or docs.
[](https://getagentictools.com/mcp/shinpr-mcp-local-rag?ref=badge)