mcp-rag-search
Provides knowledge retrieval capabilities for GitHub Copilot, enabling it to search and retrieve relevant document chunks from indexed sources.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@mcp-rag-searchsearch for LangChain integration docs"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP RAG 知识检索服务系统
面向 Claude Code、GitHub Copilot 等 Agent 工具的知识检索基础设施。基于 MCP 协议提供可插拔、可评测的知识服务。
快速开始
pip install -r requirements.txt
# 爬取文档
python main.py crawl
# 分块 + 构建索引
python main.py chunk
python main.py build
# 启动搜索 Demo
python main.py demo
# 启动 MCP 服务(供外部 Agent 调用)
python app/mcp_server.pyRelated MCP server: linked-docs
架构
用户查询 → BM25召回 + Dense召回 → RRF融合 → Reranker精排 → Top-K结果技术栈
检索: BM25 + BGE-Small Embedding + Milvus Lite
融合: Reciprocal Rank Fusion (RRF)
精排: BGE-Reranker (LoRA 微调)
协议: MCP 2.0 (stdio / SSE)
后端: FastAPI + Streamlit
消融实验结果
策略 | MRR@10 | 延迟 |
BM25 only | 0.91 | 64ms |
Dense only | 0.93 | 398ms |
Hybrid (BM25+Dense+RRF) | 0.91 ★ | 57ms |
Hybrid + Reranker | 0.73 | 4948ms |
项目结构
├── src/
│ ├── crawler/ # 文档爬虫
│ ├── indexing/ # 分块 + BM25 + Milvus
│ ├── retrieval/ # RRF融合 + Reranker
│ ├── eval/ # 消融实验 + 评测指标
│ └── finetune/ # LoRA微调
├── app/
│ ├── api.py # FastAPI 接口
│ ├── demo.py # Streamlit Demo
│ └── mcp_server.py # MCP 知识服务
└── data/
└── eval/ # 评测集This server cannot be deployed
Maintenance
Related MCP Connectors
Cloud or self-hosted knowledge for AI agents: hybrid search, reranking, GraphRAG, scoped MCP tools.
Make your knowledge agent-ready. One MCP endpoint, 5 connectors, 3 search modes.
Multi-engine search for AI agents. Trust scoring, local corpus, MCP-native. Self-hostable, BYOK.
Search your knowledge bases from any AI assistant using hybrid RAG.
Related MCP Servers
- FlicenseNot gradedqualityBmaintenanceEnables any MCP-compatible AI assistant to search, filter, and retrieve information from a local document collection using a hybrid search pipeline with vector, BM25, reranking, and LLM enrichment.4-
- FlicenseNot gradedqualityDmaintenanceEnables AI assistants to intelligently search and reference documentation using hybrid semantic + keyword search via MCP protocol.-
- FlicenseAqualityDmaintenanceProvides semantic vector search over local codebases via MCP, enabling hybrid search (dense + sparse + RRF) for any MCP client like GitHub Copilot or Claude Desktop.58-
- AlicenseNot gradedqualityBmaintenanceEnables hybrid document search (BM25 and dense) over a configurable corpus via MCP tools, returning passages and sources for AI agents to cite in answers.MIT