MCP ConfigsApr 6, 2026·2 min read

n8n MCP Server — Build Automations with AI, 1,396 Nodes

MCP server giving AI agents access to 1,396 n8n nodes and 2,709 workflow templates. Build and manage n8n automations through natural language.

TL;DR
MCP server bridging AI agents to n8n's 1,396 nodes and 2,709 workflow templates via natural language.
§01

What it is

n8n MCP Server is a bridge between AI assistants and the n8n workflow automation platform. It exposes 1,396 n8n nodes (812 core plus 584 community) and 2,709 workflow templates as MCP tools that any compatible AI agent can call. Instead of manually dragging and dropping nodes in the n8n canvas, developers describe what they want in plain English and the AI builds the workflow.

The server targets automation engineers and developers who already use n8n and want AI-assisted workflow creation. It works with Claude Code, Claude Desktop, Cursor, and Windsurf.

§02

How it saves time or tokens

Manually building an n8n workflow requires reading documentation for each node, understanding trigger types, and wiring data transformations. The MCP server pre-loads 87% documentation coverage so the AI agent can answer questions like 'What Slack actions are available in n8n?' without the user searching docs. The agent constructs valid workflow JSON directly, eliminating trial-and-error in the visual editor.

Because node documentation is served on demand through MCP tools rather than stuffed into a system prompt, token usage stays low -- the agent fetches only the docs it needs for the current task.

§03

How to use

  1. Run directly with npx (no install required):
npx n8n-mcp
  1. Or add to your MCP client config:
{
  "mcpServers": {
    "n8n-mcp": {
      "command": "npx",
      "args": ["-y", "n8n-mcp"]
    }
  }
}
  1. Restart your AI tool and start requesting automations in natural language.
§04

Example

User: Create an n8n workflow that watches a Gmail inbox for emails
      with 'invoice' in the subject, extracts the PDF attachment,
      and uploads it to Google Drive.

Agent (via n8n MCP): Here is the workflow JSON with three nodes:
  1. Gmail Trigger (filter: subject contains 'invoice')
  2. Extract attachment (binary data)
  3. Google Drive Upload (target folder: /Invoices)

The agent generates valid n8n workflow JSON that can be imported directly into your n8n instance.

§05

Related on TokRepo

§06

Common pitfalls

  • Running npx without a working n8n instance -- the MCP server generates workflow definitions, but you still need a running n8n server to execute them.
  • Expecting community nodes to have the same documentation coverage as core nodes -- 87% coverage applies to the full set, but some community nodes have minimal docs.
  • Not pinning the npx version in CI environments -- always specify a version to avoid breaking changes.

Frequently Asked Questions

Do I need a running n8n instance to use n8n MCP Server?+

The MCP server itself does not require a running n8n instance -- it generates workflow definitions and answers node documentation questions independently. However, to actually execute the workflows the AI creates, you need a running n8n instance where you import the generated JSON.

Which AI tools support n8n MCP Server?+

It works with any MCP-compatible client including Claude Code, Claude Desktop, Cursor, and Windsurf. Any tool that supports the Model Context Protocol can connect to the n8n MCP server and use its tools.

How many nodes does the server cover?+

The server provides access to 1,396 total n8n nodes: 812 core nodes maintained by the n8n team and 584 community nodes (516 of which are verified). Documentation coverage across all nodes is 87%.

Can n8n MCP Server modify existing workflows?+

Yes. The AI agent can read existing workflow JSON, understand the current node configuration, suggest modifications, and output updated workflow JSON. You then import the updated JSON into your n8n instance.

Is there a Docker option?+

Yes. You can run the server via Docker with the command: docker run -it ghcr.io/czlonkowski/n8n-mcp:latest. There is also a hosted option with a free tier offering 100 calls per day.

Citations (3)
  • n8n MCP GitHub— n8n MCP Server provides access to 1,396 nodes with 87% documentation coverage
  • n8n Official Docs— n8n is an extendable workflow automation tool with a fair-code license
  • Anthropic MCP Specification— MCP enables AI models to access external tools through a standardized protocol
🙏

Source & Thanks

Created by Romuald Członkowski. Licensed under MIT.

n8n-mcp — ⭐ 17,600+

Thank you for connecting AI agents to the n8n automation ecosystem.

Discussion

Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.