The Exercise ↔ Muscle MCP Server

Give your AI agent real-time exercise and muscle data via MCP. Structured, accurate muscle mapping for AI coaches, fitness apps, and health platforms. 850+ exercises. 63 muscles. 7,300+ mappings. MCP + REST API.

// MCP Tool Call
get_muscles_worked("bench press")

{
  "exercise": "Barbell Bench Press",
  "primary": [
    { "muscle": "Pectoralis Major (Sternal)", "group": "chest" },
    { "muscle": "Anterior Deltoid", "group": "shoulders" }
  ],
  "secondary": [
    { "muscle": "Triceps (Lateral Head)", "group": "triceps" }
  ]
}

Everything You Need

6 MCP tools for AI agents. RESTful API for everyone else. Browse, search, and analyze.

Exercise → Muscles

Query any exercise and get detailed breakdowns of primary, secondary, and stabilizer muscle groups with activation levels.

Muscle → Exercises

Find all exercises that target a specific muscle. Filter by equipment type, difficulty level, and movement pattern.

Workout Analysis

Analyze complete workouts to detect muscle imbalances, coverage gaps, and optimize programming for compound, isolation, and isometric movements.

Smart Alternatives

Get exercise substitutions ranked by muscle overlap score. Perfect for adapting workouts to equipment availability.

Exercise Instructions

630 exercises include step-by-step preparation and execution instructions. Ideal for form guidance and coaching apps.

Rich Metadata

Every exercise includes laterality (bilateral/unilateral), force type (push/pull/static), and full movement pattern classification.

Who's Building with MusclesWorked?

Used by leading fitness companies and platforms

AI Fitness Coaches

Build intelligent training programs with automatic form analysis and personalized recommendations.

Workout Trackers

Track muscle activation across every workout and visualize total body development over time.

Physical Therapy

Design targeted rehabilitation programs with confidence in anatomically accurate exercise prescriptions.

Wearable Platforms

Integrate real-time biometric data with exercise science for actionable fitness intelligence.

Content Creators

Create educational fitness content with verified, structured exercise and anatomy data.

856
Exercises
63
Muscles
7,310
Mappings
14
Muscle Groups

Built on Real Anatomy

Every exercise and muscle mapping is sourced from anatomical research and fitness science. 630 exercises include full preparation and execution instructions. Our standardized taxonomy covers 14 equipment types, 15 movement patterns, and 3 difficulty levels across compound, isolation, and isometric exercises.

Each exercise averages 8.5 muscle mappings across primary, secondary, and stabilizer roles — the most comprehensive exercise-to-muscle dataset available via API.

MCP Server + REST API

Connect your AI agent via MCP, or integrate with any stack via REST

MCP get_muscles_worked(exercise)
Get primary, secondary, and stabilizer muscles for any exercise. Works with Claude, ChatGPT, and any MCP client.
MCP find_exercises(muscle, filters...)
Find exercises targeting a muscle. Filter by equipment, difficulty, movement pattern, and type.
MCP analyze_workout(exercises[])
Analyze a workout for muscle coverage gaps, push/pull imbalances, and movement pattern diversity.
MCP get_alternatives(exercise)
Find substitute exercises ranked by muscle overlap score. Perfect for equipment constraints or variety.
MCP search_exercises(query)
Fuzzy search the exercise catalog by name. Returns matching exercises with full metadata.
MCP search_muscles(query)
Search the muscle taxonomy by name. Find muscles across all 14 anatomical groups.

All 6 tools also available as RESTful endpoints at /api/v1/ for traditional integrations. Plus browse endpoints for exercises, muscles, and muscle groups.

Simple Pricing

Start free, scale as you grow

Free
$0
Forever
  • All 856 exercises
  • All 63 muscles & 7,310 mappings
  • All 6 MCP tools
  • Full REST API access
  • Browse, search & analyze
Get Started
More tiers coming soon
MusclesWorked is free during the beta period. Paid tiers with higher rate limits will be announced later.