Anypoint MCP Server
An MCP server for MuleSoft Anypoint Platform that enables AI assistants to interact with Anypoint APIs, including accounts, API management, Exchange, monitoring, and more.
README
🔌 Anypoint MCP Server
An MCP (Model Context Protocol) server for MuleSoft Anypoint Platform — lets AI assistants like Claude talk directly to your Anypoint APIs.
🎬 Demo
✨ Features
| Module | Prefix | What it does |
|---|---|---|
| 🏢 Accounts | accounts_* |
Profile, org & environment info |
| ✏️ API Designer | api_designer_* |
Browse & manage Design Center projects |
| 📦 Exchange | exchange_* |
Search & inspect Exchange assets |
| 🔧 API Manager | api_manager_* |
API instances, SLA tiers, contracts |
| 🛡️ API Policies | api_policy_* |
Apply & manage API gateway policies |
| 📊 Metrics | metrics_* |
Traffic, latency & error rate data |
| 📡 Monitoring | monitoring_* |
ARM monitoring queries |
| 🚀 Runtime Manager | runtime_* |
CloudHub 2.0 deployments, Start/Stop/Restart |
| 📬 MQ | mq_* |
Anypoint MQ queues (sandbox by default) |
| 🔐 Secrets Manager | secrets_* |
Secrets Manager vaults & secrets |
| 📋 Reports | reports_* |
On-demand aggregated reports |
🚀 Quick Start
1. Install & build
npm install
npm run build
2. Configure credentials
Create a .env file (a blank template is already included):
ANYPOINT_CLIENT_ID=your-connected-app-client-id
ANYPOINT_CLIENT_SECRET=your-connected-app-client-secret
ANYPOINT_ORG_ID=your-organization-id # optional if profile has one org
ANYPOINT_ENV_ID=your-environment-id # default env for API/Runtime Manager tools
ANYPOINT_BASE_URL=https://anypoint.mulesoft.com # optional, this is the default
MCP_HOST=127.0.0.1
MCP_PORT=3000
MCP_PUBLIC_BASE_URL=https://your-ngrok-host.ngrok-free.dev # optional, for public UI links
💡 Use a Connected App (Client Credentials grant) from Anypoint Access Management. The server exchanges these for a Bearer token on startup and auto-renews before expiry.
3. Start the server
npm start
The server listens at:
| Endpoint | URL |
|---|---|
| 🔌 MCP (Streamable HTTP) | http://127.0.0.1:3000/mcp |
| ❤️ Health check | http://127.0.0.1:3000/health |
4. Connect your AI client
Add this to your MCP client config (e.g. Claude Desktop, Agentik):
{
"mcpServers": {
"anypoint": {
"url": "http://127.0.0.1:3000/mcp"
}
}
}
🐳 Docker
# Build
docker build -t anypoint-mcp-server .
# Run with your .env file
docker run --rm --env-file .env -p 3000:3000 anypoint-mcp-server
🛠️ Development
# Build TypeScript
npm run build
# Open MCP Inspector (browser UI to test tools)
npm run inspector
🖥️ Interactive UI
Several modules render rich interactive HTML UIs inside MCP-capable clients (like Agentik):
| UI | Tool | What you see |
|---|---|---|
| 🎨 Design Center | api_designer_list_projects |
Searchable project carousel |
| 📦 Exchange | exchange_list_assets |
Filterable asset table |
| 🔧 API Manager | api_manager_list_apis |
API instance table |
| 🚀 Runtime Manager | runtime_list_deployments |
Deployment table with Start/Stop buttons |
| 📬 MQ | mq_list_queues |
Queue list (defaults to sandbox env) |
📋 Reports
Generate on-demand aggregated reports with reports_create:
| Report type | Description |
|---|---|
runtime_flow_count |
Flow counts per deployment, grouped by environment |
api_manager_insecure_apis |
APIs with no policies or missing security policies |
metrics_api_health_summary |
Traffic, error rates & response times from Metrics API |
Reports are cached in memory for the life of the server process.
🔒 Auth & Caching
- Token is auto-renewed before expiry — no restarts needed
- Org ID and owner ID are discovered from
/accounts/api/profileon startup and cached - Pass
refresh: trueto profile/environment tools to force a cache reload - All API requests include
Authorization,x-organization-id, andx-owner-idheaders automatically
📚 API References
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。
