home / mcp / n8n-mcp mcp server

n8n-MCP MCP Server

A MCP for Claude Desktop / Claude Code / Windsurf / Cursor to build n8n workflows for you

Installation
Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "czlonkowski-n8n-mcp": {
      "command": "npx",
      "args": [
        "n8n-mcp"
      ],
      "env": {
        "MCP_MODE": "<MCP_MODE>",
        "LOG_LEVEL": "error",
        "N8N_API_KEY": "your-api-key",
        "N8N_API_URL": "https://your-n8n-instance.com",
        "WEBHOOK_SECURITY_MODE": "moderate",
        "DISABLE_CONSOLE_OUTPUT": "true"
      }
    }
  }
}

You run a Model Context Protocol (MCP) server that equips AI assistants with complete access to n8n node documentation, properties, and operations. This server helps Claude and other AI models understand 1,084 n8n nodes (537 core + 547 community), with rich schemas, operations, and real-world examples to accelerate building and validating workflows.

How to use

Connect your MCP client to the n8n-MCP server to gain structured knowledge about n8n nodes, properties, and operations. Use the available tools to search nodes, fetch detailed node data, validate configurations, and manage workflows through your n8n instance when you provide API access. Begin with discovery to find relevant templates or nodes, then validate configurations before deploying any changes. You can operate entirely through your MCP client or Claude-integrated tooling.

Typical usage patterns include: 1) searching for nodes by name or capability, 2) retrieving concise node metadata and human-readable documentation, 3) validating node configurations for correctness, and 4) managing workflows in your n8n instance via the provided management tools. Always validate thoroughly before deploying to production.

How to install

Prerequisites: Node.js is installed on your system if you plan to run a local MCP server. Docker is required for containerized runs. Ensure you have a Claude Desktop or similar MCP client configured to connect to your MCP server using stdio mode.

# Quick local setup with npx (no installation required)
npx n8n-mcp

# Basic configuration for Claude Desktop (stdio mode)
{
  "mcpServers": {
    "n8n-mcp": {
      "command": "npx",
      "args": ["n8n-mcp"],
      "env": {
        "MCP_MODE": "stdio",
        "LOG_LEVEL": "error",
        "DISABLE_CONSOLE_OUTPUT": "true"
      }
    }
  }
}

Full configuration (with n8n management tools):

{
  "mcpServers": {
    "n8n-mcp": {
      "command": "npx",
      "args": ["n8n-mcp"],
      "env": {
        "MCP_MODE": "stdio",
        "LOG_LEVEL": "error",
        "DISABLE_CONSOLE_OUTPUT": "true",
        "N8N_API_URL": "https://your-n8n-instance.com",
        "N8N_API_KEY": "your-api-key"
      }
    }
  }
}

Option B uses Docker for isolation. Pull and run the MCP image with stdio mode enabled. Example commands below show the complete flow and common environment variables.

docker pull ghcr.io/czlonkowski/n8n-mcp:latest

# Basic docker-based stdio config (documentation tools only)
{
  "mcpServers": {
    "n8n-mcp": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "--init",
        "-e", "MCP_MODE=stdio",
        "-e", "LOG_LEVEL=error",
        "-e", "DISABLE_CONSOLE_OUTPUT=true",
        "ghcr.io/czlonkowski/n8n-mcp:latest"
      ]
    }
  }
}

Full docker configuration (with n8n management tools):

{
  "mcpServers": {
    "n8n-mcp": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "--init",
        "-e", "MCP_MODE=stdio",
        "-e", "LOG_LEVEL=error",
        "-e", "DISABLE_CONSOLE_OUTPUT=true",
        "-e", "N8N_API_URL=https://your-n8n-instance.com",
        "-e", "N8N_API_KEY=your-api-key",
        "ghcr.io/czlonkowski/n8n-mcp:latest"
      ]
    }
  }
}

Additional deployment notes

If you run a local n8n instance on the same machine, you can point the MCP at the local URL http://host.docker.internal:5678 when using Docker.

Local n8n instance configuration also enables webhooks for local development by setting WEBHOOK_SECURITY_MODE to moderate. This keeps local development safe while allowing necessary webhook traffic.

Security and privacy

Protect production workflows when using AI assistance. Always work on copies, test in development, and export backups before deploying any changes.

Telemetry can be disabled if you prefer not to share anonymous usage data. Use the provided flag in your runtime configuration to opt out.

n8n integration and tools

The MCP server exposes a rich set of tools to interact with n8n nodes and workflows. You can search for nodes, fetch node details, validate configurations, and manage workflows through the included tools.

Troubleshooting and tips

Restart your MCP client after updating configuration files. If issues arise in Docker, consult the Docker Troubleshooting Guide and verify environment variables are passed correctly.

Supported environments and files

Configuration and runtime files should reflect the stdio mode unless you are using a remote HTTP endpoint. Ensure environment variables are aligned with the deployment method you choose.

Notes on usage with Claude projects

When integrating with Claude, follow the standard workflow for adding MCP servers in Claude Desktop or Claude Code. Use the stdio configuration to enable JSON-RPC messages on stdout and avoid interleaving debug logs.

Appendix: Tools exposed by the MCP server

Core tools exposed include: tools_documentation, search_nodes, get_node, validate_node, validate_workflow, search_templates, get_template, n8n_create_workflow, n8n_get_workflow, n8n_update_full_workflow, n8n_update_partial_workflow, n8n_delete_workflow, n8n_list_workflows, n8n_validate_workflow, n8n_autofix_workflow, n8n_workflow_versions, n8n_deploy_template, n8n_test_workflow, n8n_executions, and n8n_health_check.

Available tools

tools_documentation

Get documentation for MCP tools and endpoints.

search_nodes

Full-text search across all nodes and retrieve configurations and examples.

get_node

Retrieve node information with multiple detail modes and documentation options.

validate_node

Validate node configurations with minimal or full validation profiles.

validate_workflow

Validate the entire workflow for correctness and AI compatibility.

search_templates

Search templates with multiple modes including by metadata, by task, or by nodes.

get_template

Fetch full workflow templates in various modes.

n8n_create_workflow

Create a new workflow in n8n when API access is configured.

n8n_get_workflow

Retrieve workflow details in multiple modes.

n8n_update_full_workflow

Replace an entire workflow in n8n.

n8n_update_partial_workflow

Apply partial updates to a workflow using diff operations.

n8n_delete_workflow

Delete workflows from n8n.

n8n_list_workflows

List workflows with filtering and pagination.

n8n_validate_workflow

Validate a workflow in n8n by its ID.

n8n_autofix_workflow

Automatically fix common workflow errors in n8n.

n8n_workflow_versions

Manage version history and rollback for workflows.

n8n_deploy_template

Deploy templates from n8n.io to your instance with auto-fix.