Best MCP Servers for RAG & Knowledge
11 servers
MCP servers for retrieval-augmented generation, knowledge bases, vector search, and memory systems.
Context7
Up-to-date code documentation for LLMs and AI code editors.
Up-to-date code docs
For LLMs/AI editors
Community
47.0K stars
Cognee
Memory manager with graph/vector stores, 30+ data sources.
TypeScript
Local
Community
12.6K stars
Mem0
Manages coding preferences/patterns for IDEs like Cursor.
Python
Local
Community
602 stars
Graphlit
Ingests from Slack/Discord/websites into knowledge graphs.
TypeScript
Cloud
Community
373 stars
Ragie AI
Retrieves from RAG knowledge bases (e.g., Google Drive, Notion).
TypeScript
Cloud
Community
86 stars
Pinecone Assistant
Connects to Pinecone Assistant for knowledge.
Rust
Cloud
Community
41 stars
Agentset
RAG for your knowledge base. Connect your documents and data for retrieval-augmented AI agent interactions.
Tools
Official
26 stars
Neo4j Agent Memory
Memory management for AI agents using Neo4j knowledge graphs for persistent contextual memory.
Tools
Prompts
Official
15 stars
Vectorize
Advanced retrieval, deep research, anything-to-markdown file extraction, and text chunking for RAG workflows.
Tools
Resources
Official
10 stars
Frequently Asked Questions
What are MCP servers for rag & knowledge?▼
MCP (Model Context Protocol) servers for rag & knowledge allow AI applications like Claude Desktop, Cursor, and other AI clients to interact with rag & knowledge tools and services. They expose tools, resources, and prompts through the standardized MCP protocol.
How many rag & knowledge MCP servers are available?▼
There are currently 11 rag & knowledge MCP servers listed in the MCP Server Spot directory. This number grows as the community builds and submits new servers.
How do I install a rag & knowledge MCP server?▼
Most MCP servers can be installed via npm (npx) or pip (uvx). Click on any server to see its specific installation instructions and configuration snippets for your AI client.
Which AI clients support rag & knowledge MCP servers?▼
RAG & Knowledge MCP servers work with any MCP-compatible AI client, including Claude Desktop, Cursor, Windsurf, VS Code, and Zed. Each server's detail page shows verified compatible clients.