AI Assistant which answers questions with a RAG MCP and a Search Engine MCP
VerifiedAI agent answers questions using RAG and search engine MCP servers.
What this workflow does
This n8n workflow creates an AI assistant that retrieves answers from both a RAG MCP server and a search engine MCP server, combining stored data with up-to-date results through the listed integrations.
It is intended for self-hosted n8n users who need an AI agent capable of handling knowledge-base and real-time queries via MCP connections.
Who is this for?
Beginner n8n users and solo developers building AI assistants. Small teams needing quick RAG plus web search capabilities without custom coding.
What problem it solves
AI agents lack access to both internal knowledge bases and current external data. This leaves answers incomplete or outdated for real-world questions.
What it automates
Internal docs Q&A
Employee asks about company policies stored in RAG while the agent pulls latest compliance updates via search.
Research assistant
Analyst queries product features from vector DB and gets recent competitor news from the search MCP.
Customer support bot
Support agent retrieves past ticket resolutions via RAG and checks current pricing or outages through search.
How the workflow works
The 4 nodes in this automation, in order.
- 1AI Agent@n8n/n8n-nodes-langchain.agent
- 2OpenAI Chat Model@n8n/n8n-nodes-langchain.lmChatOpenAi
- 3Simple Memory@n8n/n8n-nodes-langchain.memoryBufferWindow
- 4MCP Client Tool@n8n/n8n-nodes-langchain.mcpClientTool
Apps & integrations used
How to set up AI Assistant which answers questions with a RAG MCP and a Search Engine MCP
- 1Import the workflow JSON into your self-hosted n8n instance.
- 2Deploy the separate RAG MCP Server and Search Engine MCP Server workflows.
- 3Open the MCP Client: RAG node and paste the correct SSE Endpoint URL.
- 4Open the second MCP Client node and set its SSE Endpoint to the Search MCP server.
- 5Add your OpenAI API key in the OpenAI Chat Model node credentials.
- 6Activate the workflow and test via the AI Agent chat trigger.
How to customize this workflow
- →Swap OpenAI Chat Model for another supported LLM provider.
- →Add more MCP Client Tool nodes to connect additional MCP servers.
- →Replace Simple Memory with a persistent vector store for longer conversations.
- →Change the chat trigger to a webhook for embedding in external apps.
AI Assistant which answers questions with a RAG MCP and a Search Engine MCP: pros & cons
Pros
- +Combines private RAG data with live web search in one agent
- +Uses standard MCP Client Tool nodes for easy server connections
- +Beginner-friendly structure with clear node separation
- +Works with any MCP-compatible server once endpoints are set
Cons
- –Requires self-hosted n8n and community MCP nodes
- –You must run and maintain separate MCP Server workflows
- –SSE endpoints need manual updates if servers change
Frequently asked questions
It creates an AI Agent that can query both a RAG database and a search engine through MCP servers to answer questions with internal and current data.
User reviews
Verified reviews from the community shape this listing's rating.
Loading reviews…
Similar workflows
Other ai & llm automations worth a look.