Nimbus MCP Server
Provides live user account lookup tools over HTTP for the Nimbus Furniture platform. Enables AI agents to retrieve customer account details by email or user ID via MCP.
README
🔧 Nimbus MCP Server
Standalone Model Context Protocol (MCP) server for the Nimbus Furniture platform.
Provides live user account lookup tools over HTTP, consumed by the nimbus_agent AI support agent.
What is this?
The nimbus_mcp service is a lightweight, independently deployable HTTP server that exposes two tools via FastMCP:
| Tool | Description |
|---|---|
get_user_by_email |
Look up a Nimbus customer account by email address |
get_user_account_status |
Retrieve plan type, account status, and plan details by user ID |
The agent calls these tools at runtime over HTTP using the MCP protocol — keeping live user data completely separate from the AI agent logic.
Project Structure
nimbus_mcp/
├── server.py # FastMCP server — defines and exposes MCP tools
├── data/
│ └── users.json # Nimbus customer database (8 mock users)
├── requirements.txt # Minimal dependencies (fastmcp, mcp, python-dotenv)
├── .gitignore
└── README.md
Tools Exposed
get_user_by_email(email: str) → dict
Looks up a customer by their email address.
Returns:
{
"found": true,
"id": "usr_001",
"name": "Sarah Chen",
"email": "sarah.chen@example.com",
"plan": "premium",
"account_status": "active",
"last_login": "2026-08-03T10:22:14Z",
"created_at": "2024-03-15T08:00:00Z",
"total_orders": 12,
"store_credit": 25.00
}
If not found:
{ "found": false, "error": "No Nimbus Furniture account found for email: x@x.com" }
get_user_account_status(user_id: str) → dict
Returns plan and status details for a user by ID.
Returns (Free plan):
{
"found": true,
"id": "usr_002",
"name": "James Patel",
"plan": "free",
"account_status": "active",
"plan_details": "Nimbus Free: standard shipping rates apply, White Glove Delivery available for $49/item, regular support. Upgrade to Nimbus Premium for $9.99/month..."
}
Returns (Premium plan):
{
"found": true,
"id": "usr_001",
"name": "Sarah Chen",
"plan": "premium",
"account_status": "active",
"plan_details": "Nimbus Premium: free standard shipping on all furniture orders, free White Glove Delivery & assembly on large items, 10% member discount..."
}
Returns (Suspended):
{
"found": true,
"id": "usr_004",
"name": "Derek Kim",
"plan": "free",
"account_status": "suspended",
"suspension_reason": "Payment dispute — account under review",
"plan_details": "..."
}
Setup
1. Prerequisites
- Python 3.10+
2. Create a virtual environment
python -m venv venv
# Windows
venv\Scripts\activate
# macOS / Linux
source venv/bin/activate
3. Install dependencies
pip install -r requirements.txt
4. Configure environment variables (optional)
The only configurable variable is the port. Create a .env file if needed:
MCP_SERVER_PORT=8001
The default port is 8001 if not set.
Running Locally
python server.py
The MCP server starts at:
http://127.0.0.1:8001/mcp
To verify it's running, check the console output:
INFO: Started server process
INFO: Uvicorn running on http://127.0.0.1:8001
⚠️ Always start
nimbus_mcpbeforenimbus_agent. The agent connects to this server on startup and falls back to RAG-only mode if it can't reach it.
Deploying to Render
Deploy as a Web Service from your GitHub repo:
| Setting | Value |
|---|---|
| Root Directory | nimbus_mcp |
| Build Command | pip install -r requirements.txt |
| Start Command | python server.py |
| Port | 8001 |
Environment variable to set in Render dashboard:
MCP_SERVER_PORT=8001
After deploying, copy the public URL (e.g. https://nimbus-mcp.onrender.com) and set the following in your nimbus_agent Render service:
MCP_SERVER_URL=https://nimbus-mcp.onrender.com/mcp
⚠️ Deploy
nimbus_mcpfirst, then deploynimbus_agent. The agent needs the MCP URL to be live at startup.
Updating the User Database
The user data lives in data/users.json. Each user record follows this schema:
{
"id": "usr_001",
"email": "user@example.com",
"name": "Full Name",
"plan": "premium | free",
"account_status": "active | suspended",
"last_login": "2026-08-03T10:22:14Z",
"created_at": "2024-03-15T08:00:00Z",
"total_orders": 12,
"store_credit": 25.00,
"suspension_reason": "Optional — only present if account_status is suspended"
}
After editing users.json, restart the server — the file is loaded once at startup.
Mock Users (for testing)
| Plan | Status | |
|---|---|---|
| sarah.chen@example.com | Premium | Active |
| james.patel@example.com | Free | Active |
| maria.gonzalez@example.com | Premium | Active |
| derek.kim@example.com | Free | Suspended |
| linda.wu@example.com | Premium | Active |
| tom.nguyen@example.com | Free | Active |
| aisha.johnson@example.com | Premium | Active |
| carlos.rivera@example.com | Free | Active |
Requirements
fastmcp>=2.0.0
mcp>=1.0.0
python-dotenv>=1.0.0
License
MIT — free to use, modify, and extend.
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。
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。
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。
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。
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。
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。
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。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。