McFlow

# McFlow MCP Server - AI Agent Integrations This guide shows how to integrate McFlow MCP server with various AI coding agents. ## Quick Start ```bash # Install dependencies npm install # Build the project npm run build # Start the server ./scripts/start-mcp.sh ``` ## Environment Variables - `WORKFLOWS_PATH`: Path to your n8n-workflows-template directory (default: `../n8n-workflows-template`) - `MCP_MODE`: Server mode - `stdio` or `http` (default: `stdio`) - `MCP_PORT`: Port for HTTP mode (default: `3000`) ## Integration Guides ### 1. Claude Desktop Add to your Claude Desktop configuration (`~/Library/Application Support/Claude/claude_desktop_config.json`): ```json { "mcpServers": { "mcflow": { "command": "node", "args": ["/absolute/path/to/mcflow-mcp/dist/index.js"], "env": { "WORKFLOWS_PATH": "/absolute/path/to/n8n-workflows-template" } } } } ``` ### 2. Cursor Add to your Cursor MCP configuration (`~/.cursor/mcp/config.json`): ```json { "servers": { "mcflow": { "command": "node", "args": ["./dist/index.js"], "cwd": "/path/to/mcflow-mcp", "env": { "WORKFLOWS_PATH": "../n8n-workflows-template" } } } } ``` ### 3. Windsurf Add to your Windsurf configuration (`~/.windsurf/mcp/config.json`): ```json { "mcp_servers": { "mcflow": { "type": "stdio", "command": "/path/to/mcflow-mcp/scripts/start-mcp.sh", "env": { "WORKFLOWS_PATH": "/path/to/n8n-workflows-template" } } } } ``` ### 4. Continue Add to your Continue configuration (`~/.continue/config.json`): ```json { "models": [...], "mcpServers": [ { "name": "mcflow", "command": "node", "args": ["/path/to/mcflow-mcp/dist/index.js"], "env": { "WORKFLOWS_PATH": "/path/to/n8n-workflows-template" } } ] } ``` ### 5. Cody (Sourcegraph) Add to your Cody MCP configuration (`~/.cody/mcp-servers.json`): ```json { "servers": [ { "name": "mcflow", "protocol": "stdio", "command": "node", "args": ["/path/to/mcflow-mcp/dist/index.js"], "environment": { "WORKFLOWS_PATH": "/path/to/n8n-workflows-template" } } ] } ``` ### 6. Aider For Aider, you can use the HTTP mode: ```bash # Start McFlow in HTTP mode MCP_MODE=http MCP_PORT=3000 ./scripts/start-mcp.sh # In another terminal, configure Aider export AIDER_MCP_SERVER="http://localhost:3000" aider ``` ### 7. Generic MCP Client For any MCP-compatible client, use the standard configuration: ```json { "name": "mcflow", "transport": "stdio", "command": "node", "args": ["/path/to/mcflow-mcp/dist/index.js"], "env": { "WORKFLOWS_PATH": "/path/to/n8n-workflows-template" } } ``` ## HTTP Mode (Alternative) Some AI agents may prefer HTTP mode. To enable: 1. Start the server in HTTP mode: ```bash MCP_MODE=http MCP_PORT=3000 ./scripts/start-mcp.sh ``` 2. Configure your AI agent to connect to: ``` http://localhost:3000 ``` ## Docker Support Run McFlow in a container: ```dockerfile FROM node:20-alpine WORKDIR /app COPY . . RUN npm install && npm run build ENV WORKFLOWS_PATH=/workflows CMD ["node", "dist/index.js"] ``` ```bash # Build and run docker build -t mcflow-mcp . docker run -v /path/to/n8n-workflows-template:/workflows mcflow-mcp ``` ## Troubleshooting ### Server won't start - Check that `WORKFLOWS_PATH` points to a valid directory - Ensure Node.js 18+ is installed - Run `npm install && npm run build` ### Agent can't connect - Verify the command path is absolute - Check file permissions (`chmod +x scripts/start-mcp.sh`) - Look for errors in agent logs ### Tools not working - Ensure the n8n-workflows-template directory structure is correct - Check that workflow JSON files are valid - Verify read/write permissions on workflow files ## Available Tools All AI agents will have access to these tools: - `list_workflows` - List all available n8n workflows - `read_workflow` - Read a specific workflow JSON file - `create_workflow` - Create a new workflow - `update_workflow` - Update an existing workflow - `analyze_workflow` - Analyze workflow structure and dependencies - `get_project_info` - Get project information - `validate_workflow` - Validate workflow structure - `add_node_to_workflow` - Add nodes to workflows - `connect_nodes` - Connect workflow nodes - `generate_workflow_from_template` - Generate from templates ## Support For issues or questions: - Check the [README](README.md) for basic setup - Review agent-specific documentation - Open an issue on GitHub