STEPhie MCP Server
MCP server for STEPhie tools, providing secure access to publisher data and ad forecasting via Monday.com and Google Ad Manager.
README
STEPhie MCP Server
Model Context Protocol (MCP) server for STEPhie tools, providing secure access to publisher data and ad forecasting capabilities.
Features
- 🔐 Secure Authentication via Stack Auth
- 📊 Publisher Tools - Access Monday.com publisher data
- 📈 Forecast Tools - Google Ad Manager availability forecasting
- 🚀 Fast Response - Optimized Vercel Edge Functions
- 🔄 Streaming Support - SSE for real-time updates
Available Tools
Publisher Management
getAllPublishers- Get all publishers with detailsgetPublisherFormats- Get ad formats for publishersgetPublishersByFormats- Find publishers by ad formatsfindPublisherAdUnits- Find ad units for publishersgetAdUnits- Get ad unit details
Product & Pricing
getAllProducts- Get ad productsgetAllFormats- Get ad format specificationsgetAllSizes- Get ad sizesgetAllAdPrices- Get pricing information
Forecasting
availabilityForecast- GAM availability forecasting
Setup
1. Install Dependencies
pnpm install
2. Environment Variables
Copy .env.example to .env.local and fill in your credentials:
# Stack Auth (from STEPhie)
NEXT_PUBLIC_STACK_PROJECT_ID=your_project_id
STACK_SECRET_SERVER_KEY=your_secret_key
# Monday.com API
MONDAY_API_KEY=your_monday_key
# Database (shared with STEPhie)
DATABASE_URL=your_neon_database_url
3. Local Development
pnpm dev
Server will be available at http://localhost:3000/api
Client Configuration
For Claude Desktop
Add to your Claude Desktop config:
{
"mcpServers": {
"stephie": {
"url": "https://mcp.stepnetwork.dk/api",
"headers": {
"Authorization": "Bearer YOUR_STACK_AUTH_TOKEN"
}
}
}
}
For Programmatic Access
const response = await fetch('https://mcp.stepnetwork.dk/api', {
method: 'POST',
headers: {
'Content-Type': 'application/json',
'Authorization': 'Bearer YOUR_TOKEN'
},
body: JSON.stringify({
jsonrpc: '2.0',
method: 'tools/call',
params: {
name: 'getAllPublishers',
arguments: { limit: 10 }
},
id: 1
})
});
Getting an Auth Token
- Log in to STEPhie at https://ai.stepnetwork.dk
- Navigate to Settings → API Keys
- Generate a new MCP token
- Use the token in your MCP client configuration
Deployment
Deploy to Vercel
vercel --prod
Add Domain in Vercel
- Go to Project Settings → Domains
- Add
mcp.stepnetwork.dk - Update DNS records:
Type: CNAME Name: mcp Value: cname.vercel-dns.com
API Documentation
Initialize Connection
POST /api
{
"jsonrpc": "2.0",
"method": "initialize",
"params": {
"authToken": "YOUR_TOKEN"
},
"id": 1
}
List Available Tools
POST /api
{
"jsonrpc": "2.0",
"method": "tools/list",
"id": 2
}
Call a Tool
POST /api
{
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"name": "getAllPublishers",
"arguments": {
"limit": 10,
"active": true
}
},
"id": 3
}
Development
Project Structure
stephie-mcp/
├── api/ # Vercel API endpoints
│ ├── index.ts # Main MCP endpoint
│ ├── sse.ts # SSE streaming endpoint
│ └── health.ts # Health check
├── lib/
│ ├── auth/ # Stack Auth integration
│ ├── tools/ # Tool implementations
│ ├── monday/ # Monday.com client
│ └── types/ # TypeScript types
└── package.json
Adding New Tools
- Add tool definition in
api/index.tsAVAILABLE_TOOLS array - Implement handler in
lib/tools/ - Add to executeToolHandler switch statement
Support
- Documentation: https://ai.stepnetwork.dk/docs/mcp
- Issues: https://github.com/stepnetwork/stephie-mcp/issues
- Email: support@stepnetwork.dk
License
MIT © STEP Networks
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。