n8n template

Pre-built n8n workflow that connects an AI Agent node to RAG-Forge via MCP — for AI automation agency deployments.

What you get


project/
├── workflow.json    # Import this directly into n8n
└── README.md        # Setup instructions for n8n + MCP connection

Note: unlike the other templates, the n8n template does not scaffold a Python project. It provides a workflow definition that drives an existing RAG-Forge installation through the MCP server.

Default configuration

There is no rag-forge.config.ts or pyproject.toml in this template. Configuration lives in n8n itself and in the MCP server you point it at.

The workflow exposes five MCP tools to the n8n AI Agent node:

Tool	What it does
`rag_query`	Execute a RAG query and return an answer
`rag_audit`	Run the evaluation suite against the golden set
`rag_ingest`	Index a directory of documents
`rag_inspect`	Look up a specific chunk by ID
`rag_status`	Check pipeline health

To start the MCP server that the workflow connects to:


rag-forge serve --mcp --transport http --port 3100

The workflow expects the server at http://localhost:3100/sse by default. Set the RAG_FORGE_MCP_URL credential in n8n to override this.

See the rag-forge serve and rag-forge n8n reference pages for full server options.

Recommended next steps

Start your MCP server with rag-forge serve --mcp --transport http --port 3100, then import workflow.json into n8n via Workflows → Import from File.
Set the RAG_FORGE_MCP_URL credential in n8n and configure your LLM provider in the AI Agent node.
Trigger the workflow with a test question to verify the MCP connection, then customise the agent’s system prompt for your use case.

When to upgrade

The n8n template is a parallel track to the Python-based templates (basic → hybrid → agentic → enterprise). It is not an upgrade path — it is an integration layer. For stronger retrieval or security guards, run the n8n workflow against an enterprise-template pipeline rather than switching away from this template.