VitalLink MCP

What it does

VitalLink MCP bridges a personal health data backend to any MCP-compatible AI assistant. Once connected, Claude or ChatGPT can query your real-time glucose readings, HRV, sleep quality, meal logs, skin analysis scores, and environmental conditions. The standout tool — render_health_insights — generates a full interactive health dashboard widget that renders directly inside the AI client, no browser tab needed.

Key features

• 10 health tools — glucose trends, HRV & vitals, meals & glycemic load, sleep, behavioral patterns (screen time, dopamine debt), environmental conditions (AQI, UV, pollen), skin analysis, and causal health graphs
• Live dashboard widget — render_health_insights renders an interactive chart-based dashboard as an MCP App directly inside Claude or ChatGPT
• Causal graph — discovers patterns like "high-GL meals → glucose spikes → poor sleep" and visualises the causal edges
• Cloudflare Workers backend — health data API runs serverlessly; use the default or deploy your own
• Dockerfile included — easy self-hosted deployment

Installation

git clone https://github.com/meetp06/vitalcoachai.git
cd vitalcoachai
npm install
npm run dev

Then add to your MCP client config:

{
  "mcpServers": {
    "vita-health": { "url": "http://localhost:3000/mcp" }
  }
}

Set VITA_CLOUD_URL env var to point to your own Cloudflare Worker health data API (optional — defaults to the demo worker).

Supported hosts

Claude Desktop, ChatGPT (via StreamableHTTP transport)