How do I use mcp-server-dexcom-health?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp-server-dexcom-health What's my current glucose?" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

mcp-server-dexcom-health

by IshaanRSharma

Overview Schema Source Related Servers Discussions Admin

Python

Remote

mcp-server-dexcom-health

MCP server for Dexcom CGM glucose data. Enables AI agents to access and analyze continuous glucose monitor data for health intelligence applications.

Features

Real-time glucose monitoring - Current readings with trend analysis
Historical data access - Up to 24 hours of glucose history
Time window queries - Query specific time ranges (e.g., "4-3 hours ago")
Clinical analytics - Time-in-range, GMI, CV%, AGP reports
Episode detection - Automatic hypo/hyper event identification with detailed context
Time-block analysis - Identify patterns by time of day
Persistence layer support - Pass external data for long-term analysis

Related MCP server: Dexcom MCP Server

Tools

Tool	Description
`get_current_glucose`	Current glucose reading with trend
`get_glucose_readings`	Historical readings with optional time windows
`get_statistics`	TIR, CV%, GMI, and other metrics
`get_status_summary`	Complete "how am I doing?" summary
`detect_episodes`	Find hypo/hyper episodes
`get_episode_details`	Deep analysis of each episode
`analyze_time_blocks`	Patterns by time of day
`check_alerts`	Real-time threshold alerts
`export_data`	Export for external storage
`get_agp_report`	Clinical AGP report

Installation

# Using uvx (recommended)
uvx mcp-server-dexcom-health

# Using pip
pip install mcp-server-dexcom-health

Configuration

Set environment variables:

Variable	Required	Description
`DEXCOM_USERNAME`	Yes	Dexcom username, email, or phone (+1234567890)
`DEXCOM_PASSWORD`	Yes	Dexcom password
`DEXCOM_REGION`	No	`us` (default), `ous` (outside US), or `jp` (Japan)

Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "dexcom": {
      "command": "uvx",
      "args": ["mcp-server-dexcom-health"],
      "env": {
        "DEXCOM_USERNAME": "your_username",
        "DEXCOM_PASSWORD": "your_password",
        "DEXCOM_REGION": "us"
      }
    }
  }
}

Usage Examples

Basic usage with Claude

"What's my current glucose?"

"How was my overnight control?"

"Did I have any lows today?"

"Give me my statistics for the last 12 hours"

"What about the hour before that?" (follow-up queries work!)

Time Window Queries

Query specific time ranges using start_minutes and end_minutes:

# Last 3 hours (standard)
get_glucose_readings(minutes=180)

# Specific window: 4 hours ago to 3 hours ago
get_glucose_readings(start_minutes=240, end_minutes=180)

# Stats for 6-5 hours ago
get_statistics(start_minutes=360, end_minutes=300)

# Episodes between 8-4 hours ago
detect_episodes(start_minutes=480, end_minutes=240)

Supported tools: get_glucose_readings, get_statistics, detect_episodes, export_data

Persistence Layer Integration

Tools that analyze data accept an optional data parameter for external data sources:

# Pass your own historical data
result = get_statistics(
    data=[
        {"glucose_mg_dl": 120, "timestamp": "2024-01-15T08:00:00Z"},
        {"glucose_mg_dl": 135, "timestamp": "2024-01-15T08:05:00Z"},
        # ... more readings
    ]
)

This enables building long-term analytics by storing data externally and passing it back for analysis.

Requirements

Python 3.10+
Active Dexcom Share session (requires Dexcom mobile app with Share enabled)
At least one follower configured in Dexcom Share

License

MIT

Available Tools

10 tools

analyze_time_blocksA

Analyze glucose by time of day - find when problems happen.

Breaks down data into overnight (00-06), morning (06-12), 
afternoon (12-18), and evening (18-24).

Args:
    minutes: Time period if using Dexcom API (1-1440, default 1440 = 24h)
    low: Low threshold in mg/dL (default 70)
    high: High threshold in mg/dL (default 180)
    data: Optional external readings for persistence layer integration.

Inputs: low, data, high, minutes.

check_alertsA

Check current glucose against alert thresholds.

Simple threshold check for real-time alerting.

Args:
    urgent_low: Urgent low threshold (default 54)
    low: Low threshold (default 70)
    high: High threshold (default 180)
    urgent_high: Urgent high threshold (default 250)

Inputs: low, high, urgent_low, urgent_high.

detect_episodesA

Detect hypoglycemic and hyperglycemic episodes.

Args:
    minutes: Time period if using Dexcom API (1-1440, default 1440 = 24h)
    start_minutes: Window start - minutes ago from now (e.g., 240 = 4 hours ago)
    end_minutes: Window end - minutes ago from now (e.g., 180 = 3 hours ago)
    low: Low threshold in mg/dL (default 70)
    high: High threshold in mg/dL (default 180)
    data: Optional external readings for persistence layer integration.
          Schema: [{"glucose_mg_dl": int, "timestamp": "ISO-8601"}, ...]

Examples:
    detect_episodes(minutes=180)  # Episodes in last 3 hours
    detect_episodes(start_minutes=240, end_minutes=180)  # Episodes 4h to 3h ago

Inputs: low, data, high, minutes, +2 more.

export_dataA

Export glucose readings for persistence layer integration.

Returns clean, consistent data structure for storage in external databases.
Call periodically to build long-term data history.

Args:
    minutes: Time period to export (1-1440, default 1440 = 24h)
    start_minutes: Window start - minutes ago from now (e.g., 240 = 4 hours ago)
    end_minutes: Window end - minutes ago from now (e.g., 180 = 3 hours ago)
    format: Export format - "json" or "csv" (default "json")
    data: Optional external readings to format/export instead of fetching.

Examples:
    export_data(minutes=180)  # Export last 3 hours
    export_data(start_minutes=240, end_minutes=180)  # Export 4h to 3h ago

Inputs: data, format, minutes, end_minutes, +1 more.

get_agp_reportA

Generate an Ambulatory Glucose Profile (AGP) report.

AGP is the clinical standard used by endocrinologists. Shows glucose
percentiles by time of day to identify patterns.

Args:
    minutes: Time period if using Dexcom API (1-1440, default 1440 = 24h)
    data: Optional external readings for persistence layer integration.

Inputs: data, minutes.

get_current_glucoseA

Get the current glucose reading.

Returns the most recent glucose value, trend direction, 
and timestamp. Reading must be within the last 10 minutes.

Inputs: no named parameters.

get_episode_detailsA

Get detailed context for each glucose episode - what led to it,
how severe it was, and how recovery went.

Args:
    minutes: Time period if using Dexcom API (1-1440, default 1440 = 24h)
    low: Low threshold in mg/dL (default 70)
    high: High threshold in mg/dL (default 180)
    data: Optional external readings for persistence layer integration.

Inputs: low, data, high, minutes.

get_glucose_readingsA

Get historical glucose readings.

Args:
    minutes: Number of minutes to look back from now (1-1440, default 60)
    start_minutes: Window start - minutes ago from now (e.g., 240 = 4 hours ago)
    end_minutes: Window end - minutes ago from now (e.g., 180 = 3 hours ago)
    max_count: Maximum readings to return (1-288, default 288)
    data: Optional external readings for persistence layer integration.

Examples:
    get_glucose_readings(minutes=60)  # Last hour
    get_glucose_readings(start_minutes=240, end_minutes=180)  # 4h ago to 3h ago
    get_glucose_readings(start_minutes=360, end_minutes=300)  # 6h ago to 5h ago

Note: When using start_minutes/end_minutes, we fetch from start_minutes back
and filter to the window. start_minutes should be > end_minutes (further back in time).

Inputs: data, minutes, max_count, end_minutes, +1 more.

get_statisticsA

Get glucose statistics for a time period.

Args:
    minutes: Number of minutes to analyze (1-1440, default 1440 = 24h)
    start_minutes: Window start - minutes ago from now (e.g., 240 = 4 hours ago)
    end_minutes: Window end - minutes ago from now (e.g., 180 = 3 hours ago)
    low: Low threshold in mg/dL (default 70)
    high: High threshold in mg/dL (default 180)
    data: Optional external readings for persistence layer integration.

Examples:
    get_statistics(minutes=180)  # Stats for last 3 hours
    get_statistics(start_minutes=240, end_minutes=180)  # Stats for 4h to 3h ago

Inputs: low, data, high, minutes, +2 more.

get_status_summaryA

Get a complete status summary - the "how am I doing?" tool.

Returns current glucose, recent trend, stats for the specified period,
any alerts, and a plain-English summary. Perfect for quick check-ins,
clinical dashboards, or health intelligence apps.

Args:
    minutes: Time period for context stats (1-1440, default 180 = 3 hours)

Returns:
    - current: Real-time glucose, trend, arrow
    - period_stats: Average, min, max, time-in-range for the period
    - alerts: Any recent lows/highs detected
    - summary: Plain-English interpretation with urgency level

Inputs: minutes.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

Deploy Server

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

TDQS

A4.1/5.0

Scored 2026-08-10 06:00 across 10 tools

Disambiguation4/5

Most tools target distinct functions (current, history, stats, summary, episodes, details, time analysis, alerts, export, AGP). However, get_status_summary overlaps with get_statistics and get_current_glucose, and get_episode_details overlaps with detect_episodes, though descriptions clarify the different purposes.

Naming Consistency5/5

All tool names follow a consistent verb_noun snake_case pattern (e.g., get_current_glucose, detect_episodes, analyze_time_blocks, check_alerts, export_data). No mixed conventions or vague verbs.

Tool Count5/5

10 tools is well-scoped for a CGM health analytics server. Each tool serves a distinct analytical or data-access purpose, and the count fits the typical 3-15 range comfortably.

Completeness4/5

The server covers current readings, historical data, statistics, summaries, episode detection/details, time-of-day analysis, alert checks, data export, and AGP reports. Minor gaps exist such as no tool for managing alert thresholds or user preferences, but for a read-only glucose analyzer, the coverage is strong.

Maintenance

ActivityInactive

ResponsivenessNo recent community issues

Resources

Related MCP Connectors

MedPrice AIA
io.github.medprice-ai
Hosted MCP server exposing US hospital procedure cost data to AI assistants
Biology & Medicine Finance
242
https://mcp.medprice.ai/mcp
ClawLink
dev.claw-link
MCP server connecting AI agents to 100+ apps (Gmail, Slack, Notion, GitHub) via one-click OAuth.
App Automation Workplace & Productivity
24
https://claw-link.dev/api/mcp
secureflows-mcpA
io.github.michal-lefler
MCP server for secureFlows: token-free URL builders and integration-linting tools for AI agents.
Code Analysis Developer Tools Security
181
https://www.secure-flows.com/mcp
mcpA
com.everstake
MCP server connecting AI agents to non-custodial staking data across 130+ networks.
Blockchain Cryptocurrency
1111
https://mcp.everstake.com

Related MCP Servers

Google Health MCP
Health & Wellness Fitness Tracking
davidmosiah
A
license
B
quality
B
maintenance
A local-first MCP server that enables AI agents to read user-authorized Google Health API v4 data from Fitbit, Pixel Watch, and partners via OAuth, with tokens never leaving the machine.
Updated 10 days ago (2026-11-05 16:40 UTC)
26
875 npm
84
MIT
Dexcom MCP Server
Health & Wellness Biology & Medicine
cykj40
A
license
Not graded
quality
B
maintenance
A Model Context Protocol (MCP) server that connects Claude to your personal Dexcom CGM (Continuous Glucose Monitor) for assistive diabetes management intelligence.
Updated 15 days ago (2026-10-31 10:30 UTC)
MIT
MetricBridge (formerly Health
Databases Health & Wellness
PhilipAD
A
license
A
quality
B
maintenance
Open-source MCP server for MetricBridge (formerly Health Export AI). 16 tools and 22 prompts let any MCP agent query 190 Apple Health metrics from files on your own machine. The server is read-only, has zero dependencies and makes zero network calls.
Updated 12 days ago (2026-11-03 16:50 UTC)
22
412 npm
11
MIT
Wellness CGM MCP
Health & Wellness Biology & Medicine
davidmosiah
A
license
A
quality
B
maintenance
Local-first CGM MCP for AI agents: Dexcom Developer API and FreeStyle Libre via LibreLink Up.
Updated 13 days ago (2026-11-02 17:41 UTC)
19
94 npm
1
MIT

mcp-server-dexcom-health

Features

Tools

Installation

Configuration

Claude Desktop

Usage Examples

Basic usage with Claude

Time Window Queries

Persistence Layer Integration

Requirements

License

Available Tools

Tool Schema Changelog

TDQS

Maintenance

Resources

Related MCP Connectors

Related MCP Servers

Google Health MCP

Dexcom MCP Server

MetricBridge (formerly Health

Wellness CGM MCP