Scroll to explore the system

Apple Health collects extensive biometric data but offers limited analysis tools. The data is siloed within the Health app with no way to query patterns, correlate metrics, or get AI-powered insights across the full dataset.

This Claude skill (MCP server) imports Apple Health export data and enables natural language querying. Ask questions like "How does my HRV correlate with sleep quality?" or "Show training load trend over the past 6 weeks."

The system processes workout data, heart metrics, sleep analysis, body measurements, and activity logs—fusing disparate streams into unified understanding.

01

Health Data Streams

Apple Health exports contain dozens of data types collected across multiple sensors. The skill ingests and normalizes these streams into a queryable format:

Heart Rate Variability
ms SDNN
Autonomic nervous system recovery indicator. Higher values suggest better recovery state.
Resting Heart Rate
bpm
Baseline cardiovascular fitness. Trends reveal fitness gains or stress accumulation.
Sleep Analysis
stages + duration
Core, deep, and REM sleep phases. Sleep architecture quality matters more than duration.
Active Energy
kcal
Calories burned through movement. Distinguishes training load from baseline metabolism.
Step Count
steps/day
Daily movement baseline. Non-exercise activity thermogenesis indicator.
Workouts
type + metrics
Structured exercise with heart rate zones, duration, and estimated effort.

Each stream tells a partial story. HRV alone doesn't explain why recovery is low. Sleep alone doesn't explain why energy is high. The value emerges from correlation—understanding how these signals interact over time.

Individual metrics are data points. Cross-stream correlation is insight.

02

Natural Language Analysis

Traditional health apps show dashboards. You scroll through charts looking for patterns. The cognitive load is on you to synthesize meaning from raw data.

This skill inverts the interaction model. Ask questions in natural language, get specific answers:

"How does my HRV correlate with sleep quality over the past month?"
"Show training load trend for the past 6 weeks"
"What's my average resting heart rate on days after strength training?"
"Compare my sleep efficiency on weekdays vs weekends"
"When was my best recovery week this quarter and what contributed to it?"

The system runs statistical analysis, identifies patterns, and explains findings in context. Instead of learning to read charts, you have a conversation about your health.

The interface is language. The analysis is computational. The output is understanding.

03

Privacy-First Architecture

Health data is intensely personal. Heart rate patterns, sleep disruptions, activity levels—these reveal intimate details about your life. Most health apps send this data to servers for analysis.

This skill runs entirely local. The Claude desktop app processes your health export without sending data anywhere. Your biometrics never leave your machine.

How it works: Export your data from Apple Health as a ZIP file. The skill parses the XML, builds a local database, and enables Claude to query it directly. All computation happens on your device.

Your health data stays yours. Analysis happens locally, insights stay private.

This isn't a compromise—it's a feature. Local processing means you control the data completely. No account required, no cloud sync, no terms of service. Just your data, analyzed by AI, kept on your machine.

Conversational Health Intelligence

We've been building health tools as dashboards—charts and graphs that require users to do the cognitive work of pattern recognition. The interface assumes you know what to look for.

But most people don't know what questions to ask. They don't know that HRV predicts next-day energy, or that sleep efficiency matters more than duration, or that their Tuesday workouts somehow improve Thursday recovery.

A conversational interface changes this. You can explore your health data through questions, following your curiosity rather than navigating a predetermined dashboard. The AI does the pattern matching; you do the wondering.

Health insights shouldn't require data science skills. They should require curiosity. This skill bridges that gap—your questions, your data, your understanding.