careflow-mcp
Enables Claude and other AI assistants to trigger HIPAA-compliant patient task management workflows through natural language.
README
CareFlow MCP 🏥
Production-ready healthcare workflow automation powered by n8n and the Model Context Protocol. Enables Claude and other AI assistants to trigger HIPAA-compliant patient task management workflows through natural language.
🏥 Healthcare-Ready: Includes comprehensive HIPAA compliance documentation and patient task management workflows.
Features
- 🚀 Trigger Workflows - Execute n8n workflows via webhook with custom payloads
- 📋 List Workflows - Query all active workflows from your n8n instance
- 📊 Check Status - Monitor workflow execution status in real-time
- 🏥 Healthcare-Ready - Built-in support for patient task workflows
- 🔒 Type-Safe - Full TypeScript support with Zod validation
- ⚡ Production-Ready - Comprehensive error handling and logging
- 🛠️ MCP Standard - Compatible with Claude Desktop and other MCP clients
📚 Documentation & Examples
- [Case Studies][(./case-studies/README.md)] - 5 real-world implementation examples with ROI metrics
- HEALTHCARE.md - Comprehensive HIPAA compliance guide (our differentiator!)
- Example Workflows - Importable n8n workflow JSON files
Quick Start with Examples
# 1. Import workflow to n8n
examples/healthcare-patient-task-workflow.json
# 2. Configure credentials in n8n
# 3. Ask Claude:
"Create a patient task for ID P12345 in the Patient Care workflow"
Tools Exposed
| Tool | Description | Required Params |
|---|---|---|
trigger_workflow |
Triggers an n8n workflow by name with JSON payload | workflowName, payload |
list_workflows |
Lists all active workflows from n8n | None |
get_workflow_status |
Checks execution status by ID | executionId |
create_patient_task |
Sends structured patient task to workflow | workflowName, patientId, taskType |
Prerequisites
- Node.js >= 18.0.0
- n8n instance (cloud or self-hosted) with API access
- n8n API Key (generate in n8n Settings > API)
Installation
Option 1: Via Smithery (Easiest)
Install directly from mcp.so using Smithery:
npx @smithery/cli install careflow-mcp
This will automatically:
- Install the package
- Add to your Claude Desktop config
- Prompt for required environment variables
Option 2: NPM Installation
npm install -g careflow-mcp
Option 3: From Source
# Clone the repository
git clone https://github.com/pratapsfdc22-dev/careflow-mcp.git
cd careflow-mcp
# Install dependencies
npm install
# Build the project
npm run build
# Link globally (optional)
npm link
Configuration
1. Create Environment File
cp .env.example .env
2. Configure Environment Variables
Edit .env with your n8n credentials:
# Base URL of your n8n instance
N8N_BASE_URL=https://your-n8n-instance.com
# n8n API Key (Settings > API > Create API Key)
N8N_API_KEY=n8n_api_xxxxxxxxxxxxxxxxxxxxxxxx
# Optional: Webhook Secret
N8N_WEBHOOK_SECRET=your_webhook_secret
3. Configure Claude Desktop
Add to your claude_desktop_config.json:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%/Claude/claude_desktop_config.json
{
"mcpServers": {
"n8n-workflow": {
"command": "node",
"args": [
"/path/to/careflow-mcp/dist/index.js"
],
"env": {
"N8N_BASE_URL": "https://your-n8n-instance.com",
"N8N_API_KEY": "your_api_key_here",
"N8N_WEBHOOK_SECRET": "your_webhook_secret"
}
}
}
}
Using npm global install:
{
"mcpServers": {
"n8n-workflow": {
"command": "careflow-mcp",
"env": {
"N8N_BASE_URL": "https://your-n8n-instance.com",
"N8N_API_KEY": "your_api_key_here"
}
}
}
}
Usage Examples
1. Trigger a Workflow
// Ask Claude:
"Trigger the 'Customer Onboarding' workflow with this data:
{ email: 'user@example.com', name: 'John Doe' }"
2. List All Active Workflows
// Ask Claude:
"Show me all active n8n workflows"
3. Check Workflow Execution Status
// Ask Claude:
"Check the status of execution ID: abc123"
4. Create a Patient Task
// Ask Claude:
"Create a high-priority follow-up task for patient ID P12345
in the 'Patient Care' workflow, due tomorrow"
Development
Build
npm run build
Watch Mode
npm run watch
Run Locally
npm run dev
Clean Build Artifacts
npm run clean
Project Structure
careflow-mcp/
├── src/
│ ├── index.ts # Main MCP server implementation
│ └── types.ts # TypeScript types and Zod schemas
├── dist/ # Compiled JavaScript (generated)
├── .env.example # Environment variable template
├── .gitignore # Git ignore rules
├── package.json # NPM package configuration
├── tsconfig.json # TypeScript configuration
└── README.md # This file
API Reference
trigger_workflow
Triggers an n8n workflow by name with optional JSON payload.
Input:
{
workflowName: string; // Name of the workflow
payload?: object; // Optional JSON data
}
Output:
{
"success": true,
"workflowId": "abc123",
"workflowName": "Customer Onboarding",
"response": { ... }
}
list_workflows
Lists all active workflows from n8n instance.
Input: None
Output:
{
"success": true,
"count": 5,
"workflows": [
{
"id": "abc123",
"name": "Customer Onboarding",
"active": true,
"createdAt": "2024-01-01T00:00:00.000Z",
"updatedAt": "2024-01-15T12:00:00.000Z"
}
]
}
get_workflow_status
Checks the execution status of a workflow run.
Input:
{
executionId: string; // Execution ID from trigger response
}
Output:
{
"success": true,
"execution": {
"id": "exec123",
"workflowId": "abc123",
"finished": true,
"status": "success",
"startedAt": "2024-01-15T12:00:00.000Z",
"stoppedAt": "2024-01-15T12:00:05.000Z"
}
}
create_patient_task
Sends a structured patient task to an n8n workflow.
Input:
{
workflowName: string; // Target workflow
patientId: string; // Patient identifier
taskType: string; // Task type
priority?: "low" | "medium" | "high" | "urgent";
description?: string;
dueDate?: string; // ISO 8601 format
assignedTo?: string;
metadata?: object;
}
Output:
{
"success": true,
"workflowId": "abc123",
"workflowName": "Patient Care",
"task": { ... },
"response": { ... }
}
Error Handling
The server implements comprehensive error handling with proper MCP error codes:
- Invalid Parameters -
ErrorCode.InvalidParams - Method Not Found -
ErrorCode.MethodNotFound - Internal Error -
ErrorCode.InternalError
All errors include descriptive messages for debugging.
Security Best Practices
- Never commit
.env- Always use.env.examplefor templates - Rotate API keys - Regularly update your n8n API keys
- Use webhook secrets - Add authentication to webhook triggers
- Restrict API access - Use n8n's API key permissions
- Monitor logs - Check server logs for suspicious activity
Troubleshooting
Server won't start
# Check Node.js version
node --version # Should be >= 18.0.0
# Verify environment variables
cat .env
# Check TypeScript compilation
npm run build
Workflow not found
- Verify the workflow name matches exactly (case-sensitive)
- Ensure the workflow is active in n8n
- Check API key has permission to access workflows
Authentication failed
- Verify
N8N_API_KEYis correct - Check
N8N_BASE_URLincludes protocol (https://) - Ensure API key hasn't expired
Webhook trigger fails
- Verify webhook node exists in workflow
- Check webhook path matches workflow ID
- Confirm
N8N_WEBHOOK_SECRETif required
Contributing
Contributions are welcome! Please:
- Fork the repository
- Create a feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
License
MIT License - see LICENSE file for details.
Acknowledgments
- Built with Model Context Protocol SDK
- Powered by n8n workflow automation
- Type validation by Zod
Support
- 🐛 Issues: GitHub Issues
- 💬 Discussions: GitHub Discussions
- 📖 Documentation: Full API Reference
Built with the Model Context Protocol
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。