Dexcom CGM Integration

Dexcom continuous glucose monitors (G6, G7, ONE, ONE+) provide real-time glucose data through the Dexcom Developer API.

OAuth Setup

Prerequisites

1. Register as a Dexcom developer at developer.dexcom.com 
2. Create an application in the Dexcom Developer Portal
3. Set the redirect URI to https://my.loop.health/api/cgm/dexcom/callback

Environment Variables

DEXCOM_CLIENT_ID=your-dexcom-client-id
DEXCOM_CLIENT_SECRET=your-dexcom-secret
DEXCOM_REDIRECT_URI=https://my.loop.health/api/cgm/dexcom/callback

Supported Devices

Connection Flow

Step 1: Initiate Connection

curl -X GET "https://my.loop.health/api/cgm/dexcom/connect" \
  -H "Authorization: Bearer $CLERK_JWT"

Redirects the user to the Dexcom OAuth authorization page. The getDexcomAuthUrl() function constructs the authorization URL.

Step 2: OAuth Callback

After authorization, Dexcom redirects to:

GET /api/cgm/dexcom/callback?code=AUTH_CODE&state=STATE_TOKEN

The server exchanges the authorization code for access and refresh tokens, creates a CGM connection record in the database.

Step 3: Start Receiving Data

CGM data is synced automatically via Trigger.dev every 30 minutes.

Data Collected

Glucose Readings

{
  "metricType": "glucose",
  "source": "dexcom",
  "metrics": {
    "value": 105,
    "unit": "mg/dL",
    "trend": "flat",
    "trendRate": 0.5,
    "transmitterGeneration": "G7"
  }
}

Glucose Trends

Database Storage

patient_cgm_connections

patient_glucose_readings

Automatic Sync

The syncCgmReadings Trigger.dev job runs every 30 minutes:

1. Finds all active Dexcom connections
2. Refreshes tokens if expired
3. Fetches glucose readings since last sync
4. Stores readings in patient_glucose_readings
5. Updates wearable data in Patient Graph

Glucose Ranges

Luna AI Integration

Luna AI can access CGM data through the getWearableInsights tool with metricType: 'glucose'. It provides:

• Average glucose over time periods
• Time in range (TIR) analysis
• Glucose variability (coefficient of variation)
• Trend analysis and recommendations
• Correlation with protocol adherence