This script stands on the giants of the GarminConnect PyPi as well as the InfluxDB PyPi.

• Python 3.10+
• InfluxDB 1.x instance
• Garmin Connect account

Copy .env.example to .env and fill in your credentials:

cp .env.example .env

Environment variables:

Token cache: on first login the script stores auth tokens in .garminconnect/ so subsequent runs do not need credentials.

One-shot sync for a fixed date range. Edit the start_date / end_date variables at the top of the file, or override them via CLI:

python garmin_to_influxdb.py                           # use hardcoded dates in script
python garmin_to_influxdb.py --start 2026-01-01        # override start
python garmin_to_influxdb.py --start 2026-01-01 --end 2026-03-31  # override both

Data written per day:

• Daily stats: steps, distance, calories, active minutes, floors, heart rate
• Sleep: deep, light, awake minutes
• HRV (if supported by your device)
• Intra-day step timestamps
• Activities with activityType tag (running, cycling, swimming, etc.)

Week-by-week orchestrator for backfilling historical data. Handles crashes gracefully via sync_state.json.

# Resume from last saved state (default start: 2019-04-11)
python sync_historical.py

# Dry run -- shows weeks without syncing
python sync_historical.py --dry-run

# Custom range
python sync_historical.py --start 2019-04-11 --end 2021-12-31

# Tune delays (seconds between each day, seconds between weeks)
python sync_historical.py --delay 4 --batch-delay 300

CLI flags:

Completed weeks are tracked in sync_state.json and skipped on subsequent runs. To re-sync a week, edit or delete its entry in that file.

Activities are written with an activityType tag (e.g. running, cycling, swimming) so you can filter by sport type in InfluxDB queries.

Speed fields (averageSpeed, maxSpeed) are multiplied by 3.584392177 to convert Garmin's internal unit (m/s) to km/h.

GMT offset: daily stat timestamps have 20000 seconds added to them to correct for a timezone offset. Without this correction, activities would appear on the previous day in InfluxDB.

HRV data: heart rate variability data is not available on all devices. When a device doesn't support it, the sync will log an error but continue without failing the day. It is not enabled by default.

Sync behavior: activities are fetched as a single request for the entire date range, while per-day stats (steps, sleep, heart rate, etc.) are fetched individually for each day. This means a failure in one day's per-day data does not affect the activity fetch.

An optional MCP server lives in mcp_server/. It is not required for either script above -- both work standalone.