Skip to content

INTEGRATION.md

Latest commit

 

History

History
110 lines (87 loc) · 2.92 KB

INTEGRATION.md

File metadata and controls

110 lines (87 loc) · 2.92 KB

Integration Ideas

Strava + Garmin Framework

Two data sources, different strengths:

Garmin (gccli)

  • Strengths: Health, physiology, sleep
  • Metrics: Body Battery, HRV, RHR, Sleep Score, Stress, VO2 Max
  • Frequency: Daily (automatic measurements)
  • Use case: Medical monitoring, recovery, training readiness

Strava (strava-cli)

  • Strengths: Workouts, activities, segments
  • Metrics: Distance, pace, heart rate (during activity), elevation, splits
  • Frequency: After each workout
  • Use case: Workout analysis, progress tracking, KOM hunting

Possible Scenarios

1. Recovery vs Load Correlation

# Garmin: Body Battery low
# Strava: Intense run scheduled
# → Recommendation: light recovery workout

2. Automatic Post-Workout Report

# Trigger: new activity in Strava (webhook)
# Read: average HR, pace, zones
# Read: Garmin data for the day (HRV, stress, last night's sleep)
# Generate: report "How workout affected recovery"

3. Weekly Summary

# Strava: mileage, time in zones, number of workouts
# Garmin: average Body Battery, sleep quality, RHR trend
# Result: "You ran 50 km, but slept poorly → more rest next week"

4. Training Planner

# Input: target distance (marathon/half-marathon)
# Garmin: current recovery level
# Strava: recent workouts and progress
# Output: adaptive plan accounting for physiology

Technical Implementation

Unified Data Layer

memory/workouts/
  2026-02-17-morning-run.md  # Strava activity + Garmin health

Entry format:

# Morning Run - 2026-02-17

## Strava
- Distance: 10.2 km
- Time: 54:23
- Avg Pace: 5:20/km
- Avg HR: 152 bpm
- Max HR: 168 bpm

## Garmin (same day)
- Sleep Score: 78 (7.4h)
- Body Battery: 65 → 33 (after workout)
- HRV: 45 ms
- Stress: 38 → 72 (spike during run)

## Analysis
Workout in zone 3-4, heart rate higher than usual.  
Body Battery dropped significantly - possible under-recovery from yesterday.  
Recommendation: light base run or rest tomorrow.

Script Example

#!/bin/bash
# scripts/workout_report.sh

# Get latest activity from Strava
ACTIVITY=$(strava-cli activities --limit 1 --json)

# Get Garmin data for today
HEALTH=$(gccli health today --json)

# Combine and send to AI for analysis
# → sessions_send + prompt with both JSONs
# → AI writes to memory/workouts/YYYY-MM-DD-activity.md

Roadmap

  • Strava webhook listener (auto-update on new activity)
  • Unified JSON export (strava + garmin in one file)
  • Dashboard (HTML + charts for visualizing both sources)
  • Predictive analytics (ML model on historical data)
  • Integration with clawd-coach for automatic training plan adjustments

See Also