local_faiss_mcp

概述

Local FAISS MCP Server通过模型上下文协议(MCP)提供使用FAISS进行检索增强生成(RAG)应用的本地向量数据库功能。它允许用户摄取多种格式文档(PDF、TXT、MD原生支持，pandoc可支持40多种格式)，进行分块、向量化存储，然后通过MCP工具执行语义搜索查询。服务器支持自定义嵌入模型和重排序以提高搜索相关性，适合为AI代理添加本地文档搜索功能的开发者和终端用户。

试试问 AI

装完之后，这里有 5 个你可以让 AI 做的事：

• ingest_documentdocument: string, source?: string

  Ingest a document into the vector store

  试试:Ingest the document at ./path/to/document.pdf试试:Add this text to the vector store: FAISS is a library for efficient similarity search...试试:Ingest the research paper with source 'my_paper'
• query_rag_storequery: string, top_k?: number

  Query the vector store for relevant document chunks

  试试:Search for information about FAISS similarity search试试:Find documents related to machine learning techniques试试:What does this document say about vector databases?
• extract-answerquery: string, chunks: array

  Extract the most relevant answer from retrieved document chunks with citations

  试试:Extract an answer from these document chunks about FAISS试试:Based on the retrieved documents, provide a detailed answer with citations试试:Summarize the key points from these document chunks
• summarize-documentstopic: string, chunks: array, max_length?: number

  Create a focused summary from multiple document chunks

  试试:Summarize these documents about vector databases试试:Provide a brief summary of the research papers on machine learning试试:Create a summary of the following text

可对比工具

安装

pip install local-faiss-mcp

git clone https://github.com/nonatofabio/local_faiss_mcp.git
cd local_faiss_mcp
pip install -e .

{
  "mcpServers": {
    "local-faiss-mcp": {
      "command": "local-faiss-mcp",
      "args": ["--index-dir", "/path/to/index/directory"]
    }
  }
}

FAQ

支持哪些文档格式？

原生支持TXT、MD和PDF格式。安装pandoc后，支持DOCX、HTML、EPUB和其他40多种格式。

可以使用自定义嵌入模型吗？

是的，您可以使用--embed参数在启动服务器时指定任何Hugging Face sentence-transformers模型。

local_faiss_mcp 对比