Zsadigzade/trainbud

Overview

TrainBud lets you talk to your training data through Claude, Cursor, and other MCP clients, and includes a Connect IQ watch app for on-wrist access. It is written in TypeScript and acts as a developer tool for interacting with training datasets via the Model Context Protocol.

Best for

Best for
Developers who want to query training data through MCP clients and also have a Garmin watch.

Use cases

• Query training data interactively from Claude or Cursor
• Expose training data to any MCP-compatible client
• Monitor or access training data from a Garmin Connect IQ watch

How to use

Install

npx trainbud setup

Tools exposed

• get_latest_activity
• get_activities_range
• get_sleep_data
• get_heart_rate_trends
• get_recovery_status
• get_body_composition
• get_stress_levels
• get_vo2_max_trends
• get_training_insights
• get_findings
• get_week_review
• compare_workouts
• remember_context
• get_user_context
• log_subjective
• GARMIN_EMAIL
• GARMIN_PASSWORD
• TRAINBUD_SESSION_PATH
• TRAINBUD_LOG_PATH
• TRAINBUD_CACHE_PATH

Tested with

Claude Desktop, Claude Code, Cursor, ChatGPT

Notes

TrainBud lets you talk to your training data through Claude, Cursor, and other MCP clients, and includes a Connect IQ watch app for on-wrist access. It is written in TypeScript and acts as a developer tool for interacting with training datasets via the Model Context Protocol.

0 stars on GitHub. Last updated 2026-09-08. Licensed MIT.

Use cases

• Query training data interactively from Claude or Cursor
• Expose training data to any MCP-compatible client
• Monitor or access training data from a Garmin Connect IQ watch

Pros

• Works with multiple MCP clients, not tied to one editor
• Includes a watch app for mobile access
• TypeScript codebase for easy contribution

Cons

• No stars yet, indicating very early stage or low adoption
• Watch app limits usage to Garmin devices
• Requires MCP client setup to be useful

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.