BACK_TO_FEEDAICRIER_2
Ultrahuman MCP server pipes ring data to AI agents
OPEN_SOURCE ↗
REDDIT · REDDIT// 27d agoOPENSOURCE RELEASE

Ultrahuman MCP server pipes ring data to AI agents

A community developer has released an unofficial Python MCP server wrapping the Ultrahuman Partner API, letting any MCP-compatible AI agent (Claude Code, Cursor, etc.) directly query health metrics — sleep, HRV, recovery, glucose, VO2 max — by date. The project ships with bundled skill files for morning briefs, recovery checks, weekly analytics, and a PDF report generator.

// ANALYSIS

Health wearable data piped directly into AI agent context is a genuinely useful pattern — your coding assistant knowing your recovery score before suggesting a deep-work sprint is table stakes for serious quantified-self developers.

  • Two core MCP tools cover the key use cases: a full daily dump and a single live-value fetch designed to be injected into every agent message as ambient context
  • Bundled skills add structured workflows (morning brief, weekly trends, sleep–glucose correlation, personal records) on top of the raw API wrapper
  • A prior TypeScript MCP server for Ultrahuman (Monasterolo21/Ultrahuman-MCP, June 2025) already exists with 10 stars — this Python reimplementation adds the analytics and skill layers that one lacks
  • Requires Ultrahuman Partner API access, which must be requested via in-app support — a real friction point for casual adoption
  • The broader health-data-via-MCP space is growing; Whoop has a similar community MCP server, suggesting this is becoming a pattern across wearable platforms
// TAGS
ultrahuman-mcpmcpagentapidevtoolopen-source

DISCOVERED

27d ago

2026-03-15

PUBLISHED

27d ago

2026-03-15

RELEVANCE

5/ 10

AUTHOR

Spinning-Complex