FastMCP Patient Information Server
Retrieves comprehensive patient information including personal details, insurance plans, and employee data through the get_patient tool. Supports Docker Compose deployment and integration with Watson Orchestrate for healthcare data access.
README
Horizon MCP Server
MCP server exposing get_medical_plan and get_patient tools for the Horizon healthcare platform.
Designed for deployment on horizon.prefect.io.
Files
| File | Purpose |
|---|---|
server.py |
Production server — calls live Horizon REST API |
server_mock.py |
Local dev/test — returns hard-coded sample payloads |
requirements.txt |
Python dependencies |
Dockerfile |
Container for SSE-mode deployment |
Local development
# 1. Install dependencies
pip install -r requirements.txt
# 2. Run in mock mode (stdio, no live API needed)
python server_mock.py
# 3. Or run the real server in SSE mode locally
HORIZON_BASE_URL=https://api.yourdomain.com \
HORIZON_API_KEY=your_key_here \
MCP_TRANSPORT=sse \
python server.py
Environment variables
| Variable | Default | Description |
|---|---|---|
HORIZON_BASE_URL |
https://horizon.prefect.io/api |
Base URL of the Horizon REST API |
HORIZON_API_KEY |
(empty) | Bearer token / API key for auth |
MCP_TRANSPORT |
stdio |
stdio for local, sse for HTTP hosting |
MCP_HOST |
0.0.0.0 |
Bind host (SSE mode only) |
MCP_PORT |
8000 |
Bind port (SSE mode only) |
Docker deployment
docker build -t horizon-mcp .
docker run -p 8000:8000 \
-e HORIZON_BASE_URL=https://api.yourdomain.com \
-e HORIZON_API_KEY=your_key_here \
horizon-mcp
The MCP SSE endpoint is available at:
http://localhost:8000/sse
Tools
get_medical_plan
| Argument | Type | Required | Description |
|---|---|---|---|
account_id |
str |
✅ | Account UUID |
medical_plan_id |
str |
⬜ | Full composite plan ID (e.g. ...M10412) |
auto_id |
int |
⬜ | Numeric plan key (e.g. 10412) |
Returns a full plan object with deductible, OOP, office visit, physician, facility, diagnostic, DME, emergency care, PT, mental health details, and network records.
get_patient
| Argument | Type | Required | Description |
|---|---|---|---|
contact_id |
str |
✅ | Patient contact UUID |
account_id |
str |
⬜ | Account UUID (recommended) |
Returns demographic data (name, DOB, email, phone) plus an embedded employee object with medical, dental, and vision plan summaries.
Claude Desktop / Cursor config (stdio mode)
{
"mcpServers": {
"horizon": {
"command": "python",
"args": ["/path/to/horizon_mcp/server.py"],
"env": {
"HORIZON_BASE_URL": "https://api.yourdomain.com",
"HORIZON_API_KEY": "your_key_here"
}
}
}
}
Claude Desktop / Cursor config (sse mode — after Docker deploy)
{
"mcpServers": {
"horizon": {
"url": "http://horizon.prefect.io:8000/sse"
}
}
}
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。