Summary
Your training log, answerable in a sentence.
"What am I doing today", "is my bench actually moving" and "how much have I done for back this week" are three screens in an app and one question here.
What it is
The MCP server for Arvo, an AI fitness coach. It reads your programme, workout history, personal records, body composition and caloric phase, and writes back in two places: coach memory and the caloric phase itself.
What you get
- Today and the current block: get_workout_for_day, get_active_split, get_approach_details and get_approach_history
- History and progress: get_recent_workouts, get_workout_stats, get_exercise_progress and get_personal_records
- Volume and body data: get_volume_by_muscle, get_body_progress, get_caloric_history and get_cycle_history
- Coaching context: get_active_insights, get_coach_notes, get_coach_info and get_booking_info
- get_exercise_video pulls up the form reference for a named movement
- Write access where it counts: save_memory keeps something for later, get_ai_memory reads what has been stored, and update_caloric_phase changes your phase
- get_user_profile identifies the account the rest of the answers belong to
Requirements
An Arvo account at arvo.guru and an API key created under Settings → API Keys, set as ARVO_API_KEY. It runs from npx, so there is nothing to install.
Setup effort
One command plus a key — npx -y arvo-mcp, then supply credentials