-AI-Gym-Trainer-Multi-Agent-Workout-Planner

# 🏋️‍♂️ AI Multi‑Agent Workout Pipeline A fully automated **6‑agent AI workout generation system** built using the **Google Agent Development Kit (ADK)**. This pipeline collects user workout intent, generates a workout plan, validates safety, selects equipment, saves memory, and outputs a final aggregated summary. --- ## 🚀 Features ### ✅ **1. Multi‑Agent Architecture (7 Linked Agents)** - **Decoder Agent** → Extracts: muscle, goal, training style, experience level - **Planner Agent** → Builds a personalized 6–8 exercise workout - **Equipment Agent** → Suggests required & alternative equipment - **Validator Agent** → Checks safety, volume, and difficulty - **Memory Agent** → Saves workout history (JSON) - **Recommender Agent** → Recovery, nutrition, next‑day tips - **Aggregator Agent** → Produces final clean summary for the user --- ## 🔄 Pipeline Flow ``` User Input ↓ Workout Decoder ↓ [Planner + Equipment Agents] (Parallel) ↓ Validator ↓ Memory Agent ↓ Recommender ↓ Aggregator → Final Output ``` --- ## 🧠 Memory‑Aware Workout Creation The system automatically reads your last **3 workouts** and avoids repeating exercises: - Prevents overuse injuries - Adds variation - Feels like a real personal trainer learning your pattern --- ## 🧰 Equipment Selector (Custom Tool) The system loads a JSON equipment database: ``` muscle → required_equipment + alternatives ``` If no match is found, fallback options are provided (dumbbells, bands, bodyweight, household items). --- ## 📁 Project Structure ``` project/ │ ├── agents.py ├── data/ │ └── equipments_db.json ├── memory/ │ └── workout_history.json ├── logs/ │ └── run_log.txt └── README.md Logs and Memory folder are not uploaded! ``` --- ## 📝 How to Use ### **1. Install dependencies** ```bash pip install google-adk ``` ### **2. Run the pipeline** ```bash adk run agents.py ``` The system will first display required input format. Then you can give ANY casual input like: ``` chest workout, muscle growth, supersets, intermediate ``` --- ## 📥 Example Output ``` FINAL WORKOUT SUMMARY - Muscle: chest - Goal: muscle growth - Experience: intermediate WORKOUT: - Barbell Bench Press — 4×8 — 90s - Incline DB Press — 3×10 — 75s ... EQUIPMENT: - Primary: bench, dumbbells, barbell - Alternatives: resistance bands, push‑up variations RECOMMENDATIONS: - Increase protein intake post‑workout - Stretch chest + triceps 10 minutes ... ``` --- ## 📌 Why This System Is Special - Full multi‑agent pipeline - Parallel agent execution - Memory‑aware plan creation - Logging for debugging - Clean modular architecture - Extensible for mobile apps / web UI / APIs --- ## 📌 Future Improvements - Add weekly periodisation agent - Add form‑correction agent using computer vision - Add strength‑tracking memory - Add user profile system --- ## 🤝 Contributing Pull requests are welcome! --- ## 📄 License MIT License