go-garmin

A Go client library and CLI for the Garmin Connect API.

Installation

Prerequisites

• Go 1.25 or later (install Go)
• $GOPATH/bin (usually ~/go/bin) in your PATH

Install CLI

go install github.com/llehouerou/go-garmin/cmd/garmin@latest

Verify installation:

garmin --version

Build from Source

git clone https://github.com/llehouerou/go-garmin.git
cd go-garmin
go build -o garmin ./cmd/garmin
./garmin --help

CLI Usage

Authentication

# Login to Garmin Connect
garmin login -email=your@email.com -password=yourpassword

# Logout
garmin logout

Commands

# Sleep data
garmin sleep [date]

# Wellness data
garmin wellness stress [date]
garmin wellness body-battery [date]
garmin wellness heart-rate [date]
garmin wellness spo2 [date]
garmin wellness respiration [date]
garmin wellness intensity-minutes [date]

# Activities
garmin activities list [--start=0] [--limit=20]
garmin activities get <activity-id>
garmin activities types
garmin activities splits <activity-id>
garmin activities weather <activity-id>
garmin activities details <activity-id>
garmin activities hr-zones <activity-id>
garmin activities power-zones <activity-id>
garmin activities exercise-sets <activity-id>

# Weight and HRV
garmin weight daily [date]
garmin weight range --start=YYYY-MM-DD --end=YYYY-MM-DD
garmin hrv daily [date]
garmin hrv range --start=YYYY-MM-DD --end=YYYY-MM-DD

# Training metrics
garmin metrics readiness [date]
garmin metrics vo2max [date]
garmin metrics endurance [date]
garmin metrics hill [date]
garmin metrics training-status [date]
garmin metrics load-balance [date]
garmin metrics acclimation [date]
garmin metrics race-predictions [display-name]

# Fitness age
garmin fitnessage stats --start=YYYY-MM-DD --end=YYYY-MM-DD

# Fitness stats
garmin fitnessstats get [--start=YYYY-MM-DD] [--end=YYYY-MM-DD] [--aggregation=weekly] [--metrics=calories,distance,duration]
garmin fitnessstats activities [--start=YYYY-MM-DD] [--end=YYYY-MM-DD] [--activity_type=running] [--metrics=name,startLocal,activityType]

# Biometric data
garmin biometric lactate-threshold
garmin biometric ftp
garmin biometric hr-zones
garmin biometric power-weight [date]

# Devices
garmin devices list
garmin devices settings <device-id>

# User profile
garmin profile social
garmin profile settings
garmin profile display

# Workouts
garmin workouts list [--start=0] [--limit=20]
garmin workouts get <workout-id>
garmin workouts create --file=workout.json
garmin workouts create --json='{"workoutName": "..."}'
cat workout.json | garmin workouts create
garmin workouts update <workout-id> --file=workout.json
garmin workouts delete <workout-id>
garmin workouts schedule <workout-id> <date>
garmin workouts unschedule <schedule-id>

# Exercise Library (for strength training workouts)
garmin exercises categories          # List all exercise categories
garmin exercises muscles             # List all muscle groups
garmin exercises equipment           # List all equipment types
garmin exercises list [--category=BENCH_PRESS] [--muscle=CHEST] [--equipment=DUMBBELL]
garmin exercises get <exercise-key>  # Get exercise details

# Calendar (month is 0-indexed: January=0)
garmin calendar get --year=2026 [--month=0] [--day=28] [--start=1]

All commands output JSON for easy parsing.

MCP Server (LLM Integration)

The CLI includes an MCP (Model Context Protocol) server that lets LLM assistants like Claude access your Garmin data.

Prerequisites

1. Install the CLI (see Installation)
2. Login once to create a session:

   garmin login -email=your@email.com -password=yourpassword
   

The MCP server reuses your CLI session, so you only need to login once.

Claude Code

Add to ~/.claude.json (global) or .claude/settings.json (project):

{
  "mcpServers": {
    "garmin": {
      "command": "garmin",
      "args": ["mcp"]
    }
  }
}

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "garmin": {
      "command": "garmin",
      "args": ["mcp"]
    }
  }
}

Cursor

Add to Cursor settings (Settings > MCP Servers):

{
  "garmin": {
    "command": "garmin",
    "args": ["mcp"]
  }
}

Troubleshooting

If the MCP server isn't working:

1. Verify the CLI is in your PATH: which garmin
2. Check you're logged in: garmin sleep should return data
3. Use the full path to the binary if needed:

   {
     "mcpServers": {
       "garmin": {
         "command": "/full/path/to/garmin",
         "args": ["mcp"]
       }
     }
   }
   

Available Tools

Once configured, you can ask Claude questions like:

• "How did I sleep last night?"
• "What's my training readiness today?"
• "Show me my recent activities"
• "What's my current VO2 max?"
• "How's my stress level today?"

The MCP server exposes 52 tools across these categories:

LLM-Powered Workout Creation

The MCP integration enables conversational workout creation. Ask Claude to create workouts in natural language:

Running Workouts:

"Create a 45-minute interval workout with 5-minute warmup, 6x3min at threshold with 2min recovery, and cooldown"

"Build me a long run workout in heart rate zone 2 for 90 minutes"

Strength Training Workouts:

"Create a push day workout: bench press 4x8, overhead press 3x10, tricep dips 3x12, with 90 second rest between sets"

"Build a full body strength workout targeting chest, back, and legs with compound movements"

"Create a dumbbell-only upper body workout I can do at home"

Smart Workout Planning:

"Look at my recent activities and create a recovery workout appropriate for tomorrow"

"Based on my training load, suggest and create a workout for my next session"

The LLM uses the exercise library (1,794 exercises) to find valid exercise categories and names, then constructs properly formatted workouts that sync directly to your Garmin device.

Example Conversation

You: Create a chest and triceps workout with 3 exercises each, 4 sets of 10 reps

Claude: I'll create that workout for you. Let me first find the best exercises...

[Uses list_exercises with muscle=CHEST, then muscle=TRICEPS]

I've found good options. Creating your workout now...

[Uses create_workout with the structured JSON]

Done! I've created "Chest & Triceps" workout with:

Chest:
- Barbell Bench Press: 4x10
- Incline Dumbbell Press: 4x10
- Cable Flyes: 4x10

Triceps:
- Tricep Dips: 4x10
- Skull Crushers: 4x10
- Tricep Pushdowns: 4x10

Rest: 90 seconds between sets

The workout is now in your Garmin Connect account. Would you like me to schedule it?

Workout Type Reference

The MCP tools support all Garmin workout types:

Library Usage

package main

import (
    "context"
    "fmt"
    "time"

    "github.com/llehouerou/go-garmin"
)

func main() {
    client := garmin.New(garmin.Options{})

    // Login
    err := client.Login(context.Background(), "email", "password")
    if err != nil {
        panic(err)
    }

    // Get today's sleep data
    sleep, err := client.Sleep.GetDaily(context.Background(), time.Now())
    if err != nil {
        panic(err)
    }

    fmt.Printf("Sleep score: %d\n", sleep.SleepScores.Overall.Value)
}

Architecture

This project uses a Declarative Endpoint Registry system. All endpoints are defined once in endpoint/definitions/ and automatically generate:

• CLI commands (via endpoint.CLIGenerator)
• MCP tools (via endpoint.MCPGenerator)
• Validation rules (via endpoint.Validator)

See CLAUDE.md for instructions on adding new endpoints.

License

MIT