n8n-MCP Server

n8n-MCP is a dedicated Model Context Protocol server that gives AI assistants detailed, structured access to n8n node documentation, properties, and operations. It enables models to understand and work with n8n’s extensive node catalog, including core and community nodes, with rich examples, templates, and AI tooling support. This makes building, validating, and deploying n8n workflows more reliable and faster when used in conjunction with AI assistants.

How to use

You connect your MCP client to an MCP server to enable AI-powered workflow design, validation, and execution assistance. Use the server to access node documentation, properties, operations, templates, and real-world examples. The server exposes a range of tools that your MCP client can invoke to discover templates, inspect node details, validate configurations, and manage workflows. Start by choosing a connection method that fits your environment, then configure your client to point at the MCP server and enable any necessary API credentials.

How to install

Prerequisites: Node.js on your system for local setups or Docker if you prefer containerized execution.

{
  "mcpServers": {
    "hosted_mcp": {
      "type": "http",
      "url": "https://dashboard.n8n-mcp.com",
      "args": []
    }
  }
}

Option A: Hosted service (no setup) — connect your MCP client to the hosted MCP dashboard at dashboard.n8n-mcp.com, obtain your API key, and attach it to your MCP client configuration.

Option B: Local quick start using npx (stdio communication) — install and run the MCP server locally via npx.

npx n8n-mcp

Configure your client to run in stdio mode by supplying the following environment variables. This ensures JSON-RPC messages flow correctly to stdout without extra debug logs.

{
  "mcpServers": {
    "n8n-mcp": {
      "command": "npx",
      "args": ["n8n-mcp"],
      "env": {
        "MCP_MODE": "stdio",
        "LOG_LEVEL": "error",
        "DISABLE_CONSOLE_OUTPUT": "true"
      }
    }
  }
}

Additional installation options

Option C: Local n8n-MCP installation (development) — clone the repository, install dependencies, build, and run in stdio mode or HTTP mode for remote access.

# 1. Clone the repository
git clone https://github.com/czlonkowski/n8n-mcp.git
cd n8n-mcp

# 2. Install dependencies and build
npm install
npm run build
npm run rebuild

# 3. Start in stdio mode for Claude Desktop
npm start

# Optional: start in HTTP mode for remote access
npm run start:http

Notes on configuration and security

You will typically set MCP_MODE to stdio for local Claude Desktop usage. If you are connecting to a local or remote n8n instance, provide N8N_API_URL and (optionally) N8N_API_KEY. For local development, ensure localhost webhooks are allowed by configuring WEBHOOK_SECURITY_MODE appropriately.

Configuration files for Claude Desktop should be updated when you change MCP connection details. After updating, restart Claude Desktop to apply changes.

Security and privacy considerations

Respect safety practices: never modify production workflows using AI directly. Always create a copy, test in a development environment, and export backups before deploying. Telemetry and privacy settings can be controlled per deployment.

Troubleshooting and tips

If you encounter issues with Docker, consult the Docker Troubleshooting Guide. When running locally, ensure the MCP stdio channel remains the only JSON-RPC pathway to avoid mixed debug logs.

Quick reference: common deployment options

Hosted service: dashboard.n8n-mcp.com for immediate access. Local quick start: npx n8n-mcp with MCP_MODE set to stdio. Local development: clone, build, run node-based start in stdio mode.

Additional notes

For Claude Desktop usage, include the MCP connection in your Claude configuration and restart the client after changes.

Available tools

tools_documentation

Get documentation for MCP tools and endpoints to guide integrations.

search_nodes

Full-text search across n8n nodes with optional examples and source filtering.

get_node

Retrieve unified node information with multiple modes such as standard, docs, and properties search.

validate_node

Validate node configurations quickly or comprehensively before deployment.

validate_workflow

Validate entire workflows, including AI agent considerations.

search_templates

Search workflow templates by metadata, task, or node composition.

get_template

Fetch complete workflow JSON from templates in different modes.

n8n_create_workflow

Create new workflows in your n8n instance.

n8n_update_partial_workflow

Apply diff-based updates to existing workflows.

n8n_delete_workflow

Delete workflows from your n8n instance.

n8n_validate_workflow

Validate a workflow inside n8n after changes.

n8n_autofix_workflow

Automatically fix common workflow issues.

n8n_health_check

Check connectivity and capabilities of the n8n API.