n8n MCP Server

Connect your AI assistant to n8n workflow automation platform through the Model Context Protocol. This server provides direct access to n8n's entire REST API, enabling AI-powered workflow management, execution monitoring, and automation control.
Quick Start
- Install with uvx:
uvx n8n-mcp-server
- Add to your Claude Desktop config:
{
"mcpServers": {
"n8n": {
"command": "uvx",
"args": ["n8n-mcp-server"],
"env": {
"N8N_HOST": "https://your-n8n.com",
"N8N_API_KEY": "your-api-key"
}
}
}
}
- Restart Claude Desktop and start automating!
Features
- 40+ Auto-generated Tools - Full access to n8n's REST API
- Smart JSON Handling - Automatically fixes JSON serialization issues
- Custom Lightweight Tools - Optimized tools for working within token limits
- Full Authentication - Secure API key authentication
- Built with FastMCP - Reliable, high-performance MCP implementation
Custom Tools for Large Workflows
The server includes custom tools designed for handling large workflows that might exceed token limits:
list_workflows_minimal
- Lists workflows with only essential metadata (id, name, active, dates, tags)
get_workflow_summary
- Gets workflow info with node/connection counts instead of full data
partial_update_workflow
- Updates specific nodes without sending the entire workflow
add_nodes_to_workflow
- Adds new nodes and automatically handles connection rewiring
File-Based Workflow Management
For working with complete workflows, use these file-based tools:
create_workflow_from_file
- Create a new workflow from a JSON file
update_workflow_from_file
- Update an existing workflow from a JSON file
export_workflow_to_file
- Export a workflow to a JSON file for backup or sharing
These tools solve token limit issues by working with files instead of passing large JSON structures.
Prerequisites
- Python 3.8 or higher
- An n8n instance with API access enabled
- n8n API key
Installation
Using uvx (Recommended)
The easiest way to use n8n MCP Server is with uvx, which runs the server in an isolated environment:
{
"mcpServers": {
"n8n": {
"command": "uvx",
"args": ["n8n-mcp-server"],
"env": {
"N8N_HOST": "https://your-n8n-instance.com",
"N8N_API_KEY": "your-api-key-here"
}
}
}
}
Using pip
pip install n8n-mcp-server
From Source
git clone https://github.com/andrewlwn77/n8n-mcp-server.git
cd n8n-mcp-server
- Install in development mode:
pip install -e .
Configuration
The server requires two environment variables:
You can provide these through:
- Environment variables in your MCP client configuration (recommended for uvx)
- A
.env
file in your working directory:
N8N_HOST=https://your-n8n-instance.com
N8N_API_KEY=your-api-key-here
Usage
Running the Server
Start the MCP server:
n8n-mcp-server
Or if running from source:
python -m n8n_mcp
The server will:
- Connect to your n8n instance
- Fetch the OpenAPI specification
- Generate MCP tools for all available endpoints
- Start listening for MCP requests
What Can You Do?
With n8n MCP Server, your AI assistant can:
- Manage Workflows - Create, update, delete, and organize automation workflows
- Execute Workflows - Trigger workflow runs and pass custom data
- Monitor Executions - Check workflow status, review logs, and handle errors
- Work with Credentials - Safely manage authentication for external services
- Handle Large Workflows - Use optimized tools designed for token limits
- Update Specific Nodes - Modify individual workflow nodes without affecting others
- Add Nodes Dynamically - Insert new nodes and automatically rewire connections
- Import/Export Workflows - Work with workflow files for easy sharing and backup
Recommended Workflow Management: Use the file-based tools (create_workflow_from_file
, update_workflow_from_file
, export_workflow_to_file
) for managing complete workflows. These tools avoid token limits by working with JSON files on your filesystem.
Example commands you can give your AI assistant:
- "List all my active workflows"
- "Execute the 'Daily Report' workflow with today's date"
- "Show me failed executions from the last 24 hours"
- "Create a new workflow from /path/to/workflow.json"
- "Export workflow abc123 to backup.json"
- "Update workflow xyz789 from updated-workflow.json"
- "Add a Slack notification node to my error handling workflow"
- "Update the schedule trigger to run every hour instead of daily"
MCP Client Configuration
Claude Desktop
Add to your Claude Desktop configuration:
Using uvx (Recommended)
{
"mcpServers": {
"n8n": {
"command": "uvx",
"args": ["n8n-mcp-server"],
"env": {
"N8N_HOST": "https://your-n8n-instance.com",
"N8N_API_KEY": "your-api-key-here"
}
}
}
}
Using pip installation
{
"mcpServers": {
"n8n": {
"command": "n8n-mcp-server",
"env": {
"N8N_HOST": "https://your-n8n-instance.com",
"N8N_API_KEY": "your-api-key-here"
}
}
}
}
Using .env file
If you have a .env
file in your working directory:
{
"mcpServers": {
"n8n": {
"command": "n8n-mcp-server"
}
}
}
Security Notes
- Keep your
.env
file secure and never commit it to version control
- Use environment-specific API keys with minimal required permissions
- Consider using read-only API keys for development/testing
Troubleshooting
No Tools Showing in MCP Client
- Ensure the server started successfully
- Check that your n8n credentials are correct
- Verify the MCP client can connect to the server
Connection Failed
- Verify your n8n instance URL is correct and includes the protocol (https://)
- Check that your API key is valid and has the necessary permissions
- Ensure your n8n instance has API access enabled
- Make sure the n8n instance is accessible from your network
Missing Tools
- The available tools depend on your n8n instance version and configuration
- Some endpoints may require admin permissions
- Check the server logs for any errors during initialization
Related Documentation
Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
License
MIT License - see LICENSE file for details.