mcp-n8n-server

# MCP n8n API Server An mcp server that provides an interface to interact with n8n workflows through the Model Context Protocol (MCP). ## Features - List all n8n workflows - Trigger specific workflows with custom data - Integration with Claude Desktop and other MCP clients ## Installation ### Global Installation (Recommended for Remote Usage) ```bash npm install -g @ahmad.soliman/mcp-n8n-server ``` Then configure your n8n connection: 1. Create a `.env` file in your working directory 2. Add your n8n API information (see Configuration section below) ### Using with npx (No Installation Required) You can run the server directly with npx: ```bash npx -y @ahmad.soliman/mcp-n8n-server ``` ### Local Installation ```bash git clone https://github.com/ahmadsoliman/mcp-n8n-server.git cd mcp-n8n-server npm install ``` ## Configuration Create a `.env` file with the following variables: ``` # n8n Host URL (required) N8N_HOST_URL=https://your-n8n-instance.com # n8n Project ID (optional - only needed for cloud instances) PROJECT_ID=your_project_id_here # n8n API Key (required) N8N_API_KEY=your_api_key_here ``` ## Usage ### As a Remote MCP Server After installing globally, you can use it as a remote MCP server with Claude AI: 1. Configure Claude AI to use this as a remote MCP server using the following JSON configuration: ```json { "mcpServers": { "n8n": { "command": "npx", "args": ["-y", "@ahmad.soliman/mcp-n8n-server"], "env": { "N8N_HOST_URL": "", "PROJECT_ID": "", "N8N_API_KEY": "" } } } } ``` 2. Add the following to your prompt or instructions to Claude: ``` You have access to a remote MCP server for n8n integration. Use it to: - List all n8n workflows - Trigger webhooks and workflows - Get information about available webhooks ``` ### As a Local MCP Server You can run the server locally and connect to it from Claude Desktop: ```bash # Start the server npm start ``` Then configure Claude Desktop to use this MCP server: Edit `~/Library/Application Support/Claude/claude_desktop_config.json`: ```json { "mcpServers": { "n8n": { "command": "npx", "args": ["-y", "@ahmad.soliman/mcp-n8n-server"], "env": { "N8N_HOST_URL": "", "PROJECT_ID": "", "N8N_API_KEY": "" } } } } ``` Alternatively, if you've cloned the repository locally: ```json { "mcpServers": { "n8n-server": { "command": "node", "args": ["/ABSOLUTE/PATH/TO/server-n8n/build/index.js"] } } } ``` ## Setup 1. Install dependencies: ```bash npm install ``` 2. Configure environment variables: - Copy `.env.example` to `.env` (if not already done) - Update the following variables in `.env`: - `N8N_API_URL`: Your n8n instance URL (default: http://localhost:5678) - `N8N_API_KEY`: Your n8n API key 3. Start the server: ```bash # Start the MCP server (for integration with Claude Desktop and other MCP clients) npm run mcp ``` For development with auto-reload: ```bash npm run dev ``` ## MCP Tools (for LLM Integration) The MCP server exposes the following tools for use with Claude Desktop or other MCP clients: ### List Workflows The `list-workflows` tool returns a list of all available n8n workflows. ### List Workflow Webhooks The `list-workflow-webhooks` tool returns all webhooks from a specific workflow. Parameters: - `id`: The ID of the workflow to get webhooks from ### Call Webhook (GET) The `call-webhook-get` tool allows calling a webhook with a GET request. Parameters: - `url`: The webhook URL to call ### Call Webhook (POST) The `call-webhook-post` tool allows calling a webhook with a POST request. Parameters: - `url`: The webhook URL to call - `data`: Data to send in the POST request body