
A fitness-focused MCP that surfaces three core tools: `create_workout` for generating training plans, `track_progress` for logging sessions and metrics, and `suggest_exercises` for pulling exercise recommendations. Built by MEOK AI Labs, it also calculates training volume across workouts. You'd reach for this when building AI agents that need to construct periodized programs, monitor client adherence, or dynamically recommend movements based on goals or equipment. Ships via pip or uvx, runs over stdio, and pairs with Claude Desktop out of the box. The repo highlights EU AI Act compliance hooks, though the fitness domain itself sits outside high-risk categories. Solid starting point if you're prototyping coaching bots or workout logging interfaces and want structured access to exercise primitives without rolling your own schema.