Dexcom G7 MCP Server
Provides access to Dexcom G7 continuous glucose monitor data, enabling AI assistants to retrieve current glucose readings and historical data.
README
Dexcom G7 MCP Server
A Model Context Protocol (MCP) server that provides access to Dexcom G7 continuous glucose monitor data. This server allows AI assistants and other MCP clients to retrieve current glucose readings and historical data from your Dexcom G7 device.
Features
- Current Glucose Reading: Get the latest glucose value with trend information
- Historical Data: Retrieve glucose readings for a specified time period
- Dual Units: Values provided in both mg/dL and mmol/L
- HTTP Transport: Native HTTP/JSON-RPC interface for easy integration
- MCP Compatible: Full Model Context Protocol compliance
Quick Start
Using Docker
# Build the image
docker build -t dexcom-mcp .
# Run with your credentials
docker run -p 8007:8007 \
-e DEXCOM_USERNAME="your-dexcom-username" \
-e DEXCOM_PASSWORD="your-dexcom-password" \
-e DEXCOM_REGION="us" \
dexcom-mcp
Using MCP-Compose
For orchestrated deployment with other MCP servers, use mcp-compose:
# mcp-compose.yaml
version: '1'
servers:
dexcom:
image: dexcom-mcp:local
runtime: docker
build:
context: .
dockerfile: Dockerfile
protocol: http
http_port: 8007
env:
HTTP_PORT: "8007"
DEXCOM_REGION: "us"
DEXCOM_USERNAME: "your-dexcom-username"
DEXCOM_PASSWORD: "your-dexcom-password"
capabilities: [tools]
networks: [mcp-net]
Then start with:
mcp-compose up dexcom
Environment Variables
| Variable | Description | Default | Required |
|---|---|---|---|
DEXCOM_USERNAME |
Your Dexcom account username | - | Yes |
DEXCOM_PASSWORD |
Your Dexcom account password | - | Yes |
DEXCOM_REGION |
Dexcom region (us or ous) |
us |
No |
HTTP_PORT |
HTTP server port | 8007 |
No |
Note: For users outside the US, set DEXCOM_REGION=ous (Outside US).
Available Tools
get_current_glucose
Retrieves the most recent glucose reading from your Dexcom G7.
Parameters: None
Example Response:
🩸 Current Glucose: 120 mg/dL (6.66 mmol/L)
📈 Trend: Steady
⏰ Time: 2024-01-15 14:30:00
get_glucose_history
Retrieves historical glucose readings for a specified time period.
Parameters:
hours(integer, optional): Number of hours of history to retrieve (default: 6)
Example Response:
📊 Last 6h glucose readings:
1. 2024-01-15 14:30:00 - 120 mg/dL (6.66 mmol/L) [Steady]
2. 2024-01-15 14:25:00 - 118 mg/dL (6.55 mmol/L) [Steady]
3. 2024-01-15 14:20:00 - 115 mg/dL (6.39 mmol/L) [Slowly Rising]
...
API Usage
Initialize Connection
curl -X POST http://localhost:8007/ \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 1,
"method": "initialize",
"params": {
"protocolVersion": "2024-11-05",
"capabilities": {},
"clientInfo": {"name": "test-client", "version": "1.0.0"}
}
}'
Get Current Glucose
curl -X POST http://localhost:8007/ \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 2,
"method": "tools/call",
"params": {
"name": "get_current_glucose",
"arguments": {}
}
}'
Get Glucose History
curl -X POST http://localhost:8007/ \
-H "Content-Type: application/json" \
-d '{
"jsonrpc": "2.0",
"id": 3,
"method": "tools/call",
"params": {
"name": "get_glucose_history",
"arguments": {"hours": 12}
}
}'
Client Integration
Claude Desktop
Add to your Claude Desktop MCP settings:
{
"servers": [
{
"name": "dexcom",
"httpEndpoint": "http://localhost:8007",
"capabilities": ["tools"],
"description": "Dexcom G7 glucose monitor"
}
]
}
OpenWebUI
Use the auto-generated OpenAPI specification at http://localhost:8007/openapi.json (when using mcp-compose proxy).
Architecture
Transport Protocol
This server uses HTTP transport instead of the traditional STDIO transport used by many MCP servers. This provides several advantages:
- Direct HTTP/JSON-RPC communication
- No need for stdio redirection or process management
- Better error handling and connection management
- Native support for web clients and proxies
Socat Integration (Optional)
The included entrypoint.sh provides socat integration for environments that expect STDIO transport. When used with MCP-Compose's socat hosting feature, it bridges between TCP sockets and the server process:
# The entrypoint wraps the server for STDIO compatibility
socat TCP-LISTEN:12345,reuseaddr,bind=0.0.0.0 EXEC:"python server.py"
This allows the HTTP-based server to work with STDIO-expecting orchestrators while maintaining the benefits of HTTP transport.
For more information about socat hosting and orchestration, see the mcp-compose documentation.
Dependencies
- pydexcom: For Dexcom API access
- FastAPI: HTTP server framework
- uvicorn: ASGI server
- mcp: Model Context Protocol support
Requirements
- Python 3.11+
- Valid Dexcom account credentials
- Active Dexcom G7 device with data sharing enabled
Security Considerations
- Store credentials securely using environment variables
- Consider using Docker secrets for production deployments
- Restrict network access to trusted clients
- Monitor access logs for unauthorized usage
Troubleshooting
Common Issues
Authentication Failed
- Verify your Dexcom username and password are correct
- Ensure you can log into the Dexcom mobile app
- Check that data sharing is enabled in your Dexcom account
No Recent Data
- Confirm your G7 sensor is active and transmitting
- Check that your phone/receiver has recent connectivity to Dexcom servers
- Verify the sensor hasn't expired
Connection Timeouts
- Dexcom servers occasionally experience delays
- Retry requests after a brief wait
- Check your internet connectivity
Debug Mode
Enable debug logging by setting the log level:
docker run -p 8007:8007 \
-e DEXCOM_USERNAME="your-username" \
-e DEXCOM_PASSWORD="your-password" \
-e PYTHONUNBUFFERED=1 \
dexcom-mcp
Disclaimer
This software is not affiliated with or endorsed by Dexcom, Inc. Use of this software requires valid Dexcom account credentials and is subject to Dexcom's terms of service. This tool is for personal use and should not be used as a substitute for proper medical monitoring and care.
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。