McFlow
Enables creation, management, and deployment of n8n workflows with enhanced context and automation capabilities. Features code extraction, automatic documentation, template generation, and intelligent project organization for streamlined workflow development.
README
McFlow - MCP Server for n8n Workflow Automation
McFlow is a Model Context Protocol (MCP) server that provides enhanced context and automation capabilities for creating and managing n8n workflows. It streamlines workflow development with code extraction, automatic documentation, and intelligent project organization.
Features
- Workflow Management: Create, read, update, and deploy n8n workflows
- Code Extraction & Compilation: Maintain code in separate files with automatic injection during deployment
- Automatic Deployment: Smart compilation and deployment with code injection
- Template Generation: Built-in templates for common workflow patterns
- Multi-Project Support: Handle simple or complex repository structures
- Workflow Analysis: Analyze dependencies, validate structure, and optimize performance
- Auto-Documentation: Automatically maintain workflow documentation
Quick Start
Prerequisites
- Node.js 18+ installed
- n8n instance running (local or remote)
- MCP-compatible client (Claude Desktop, Continue.dev, etc.)
Installation
# Clone the repository
git clone https://github.com/mckinleymedia/mcflow-mcp.git
cd mcflow-mcp
# Install dependencies
npm install
# Build the server
npm run build
# Optional: Quick setup script
./scripts/setup-agent.sh
Configuration
For Claude Desktop
Add to your Claude Desktop configuration (~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": {
"mcflow": {
"command": "node",
"args": ["/absolute/path/to/mcflow-mcp/dist/index.js"],
"env": {
"WORKFLOWS_PATH": "/path/to/your/n8n-workflows"
}
}
}
}
For Continue.dev
Add to your Continue configuration:
{
"models": [
{
"model": "claude-3-5-sonnet",
"mcpServers": {
"mcflow": {
"command": "node",
"args": ["/absolute/path/to/mcflow-mcp/dist/index.js"],
"env": {
"WORKFLOWS_PATH": "/path/to/your/n8n-workflows"
}
}
}
}
]
}
Basic Usage
Once configured, McFlow provides these commands through your MCP client:
# List all workflows
mcflow list
# Create a new workflow
mcflow create --name "my-automation" --workflow {...}
# Extract code from nodes into separate files
mcflow extract
# Compile workflows (inject code from files)
mcflow compile --output
# Deploy workflows to n8n (with automatic compilation)
mcflow deploy
# Generate from template
mcflow generate --template webhook-api --name "api-handler"
Key Concepts
Code Extraction & Compilation System
McFlow separates code from workflow JSON for better development:
- Extract: Pull code/SQL/prompts from nodes into
workflows/nodes/directory - Edit: Use your IDE with full syntax highlighting and tooling
- Compile: Process workflows to inject code from external files
- Deploy: Automatically compile and deploy to n8n
Project Structure
McFlow adapts to your repository structure:
# Simple structure
your-project/
├── workflows/
│ ├── flows/ # Workflow JSON files
│ └── nodes/ # Extracted code files
# Multi-project structure
your-repo/
├── project1/
│ └── workflows/
├── project2/
│ └── workflows/
Workflow Templates
Built-in templates for common patterns:
webhook-api- REST API endpoint workflowscheduled-report- Scheduled data reportsdata-sync- Database synchronizationerror-handler- Error handling workflowapproval-flow- Multi-step approval process
Available Tools
Core Workflow Operations
list- List all workflows in projectread- Read workflow JSONcreate- Create new workflowupdate- Update existing workflowdelete- Remove workflow
Deployment & Execution
deploy- Deploy workflows to n8n (with code injection)export- Export workflows from n8nexecute- Run a workflowdeployed- List deployed workflowsactivate- Activate/deactivate workflows
Code Management
extract_code- Extract node content to fileslist_code- List extracted code filescreate_module- Create shared code modules
Analysis & Validation
analyze- Analyze workflow structurevalidate- Check for common issuesstatus- Show deployment statuscredentials- Analyze credential requirements
Development
# Run in development mode
npm run dev
# Run tests
npm test
# Build for production
npm run build
Troubleshooting
Common Issues
- Workflows not deploying: Ensure you're using
mcflow deploy, not n8n CLI directly - Code not injecting: Check that extracted files exist in
workflows/nodes/ - Connection errors: Verify n8n is running and accessible
- Missing credentials: Use
mcflow credentials analyzeto check requirements
Debug Mode
Enable debug logging:
export DEBUG=mcflow:*
npm run dev
Documentation
Getting Started
- Overview - Introduction to McFlow concepts and benefits
- Architecture - Technical architecture and design patterns
- Integration Guide - Setting up with different MCP clients
Development
- Workflow Compilation - Code extraction and compilation system
- Node Reference - Real n8n nodes reference and best practices
- Git Workflow - Version control guidelines and best practices
- Troubleshooting - Common issues and solutions
Operations
- Credentials - Credential security and management
- Common Issues - n8n-specific issues and fixes
AI Agent Documentation
- AI Instructions - Critical instructions for AI agents
- Workflow Guidelines - Workflow creation instructions for AI
- LLM Configuration - LLM parameters for AI nodes
Contributing
Contributions are welcome! Please read our contributing guidelines and submit PRs to the main repository.
License
MIT License - see LICENSE file for details
Support
- GitHub Issues: mcflow-mcp/issues
- Documentation: Check the
docs/folder for detailed guides
For AI Agents: See docs/ai/instructions.md for critical usage instructions
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。