Local FAISS MCP Server

Example Use Cases

• 1

  Local vector storage and semantic search without external dependencies.

• 2

  Ingest and query documents in various formats including PDF, TXT, MD, and more with pandoc support.

• 3

  Enhanced search relevance using customizable embedding and re-ranking models.

Description

This MCP server enables efficient local semantic search by ingesting, chunking, and embedding documents into a FAISS vector store. It supports multiple document formats, persistent storage, and customizable embedding and re-ranking models. The server integrates seamlessly with MCP-compatible AI agents, allowing them to perform natural language queries over local document collections with citation-backed answers and summaries.