🏋️‍♂️ Trika.ai — Unified Fitness Intelligence Monorepo

This repository serves as the central monorepo for the Trika.ai Ecosystem containing:

Component	Description
trika_backend	Backend API powering workout tracking, AI inference, and authentication
trika_dashboard	Web dashboard for users, coaches, and AI-enhanced analytics
demovideos.zip	Demo workout files used for AI Workout Classification Model training and evaluation (Git LFS)

Trika.ai combines Computer Vision, Deep Learning, and Agentic RAG to deliver a smart fitness experience, including automatic workout detection, pose correctness scoring, and personalized fitness recommendations.

---

📂 Repository Structure

Trika-MONOREPO/
│
├── trika_backend/         # Submodule: Backend services
├── trika_dashboard/ 
├── screenshots/       # Submodule: Dashboard UI
├── demovideos.zip         # Stored via Git LFS
├── .gitattributes
├── .gitmodules
└── README.md

---

🧠 Core Features

• Real-time Workout Classification (Squats, Push-ups, Lunges, Planks, etc.)
• Pose Accuracy Scoring using MediaPipe/BlazePose
• Dashboard with Progress Analytics & History
• Coach Mode: AI-driven posture corrections
• Content system for training programs & personalized guidance
• Optimized inference engine for low-latency evaluation

---

🤖 Workout Classification Model

Model architecture summary:

• Input: Video frames or real-time webcam feed
• Feature extraction: Pose keypoints (MediaPipe/BlazePose)
• Temporal modeling: CNN + BiLSTM hybrid architecture
• Output: Workout label + confidence + rep count

Model Inference Flow

Video → Pose Extraction → Keypoint Array → Model → Classification + Reps + Form Score

---

📸 Screenshots & Demo Media

Replace these placeholders with your actual files.

User Flow

🔷 Dashboard — User Progress & Session History

🔷 Live Workout Classification — Web Interface

🔷 AI Coach Feedback — Rep Counter & Form Evaluation

Trika Bot / Trainer

Trika Soul / Mind

Challenge AI custum generated (Personalized)

Habbit And Schedule (Trika Tracker)

---

🎥 Workout Classification Demo Video

Embed GIF/MP4 later:

Or link a YouTube demo:

🔗 Demo: Add YouTube link here

---

🚀 Getting Started

1️⃣ Clone with Submodules

git clone --recurse-submodules https://github.com/kelvinprabhu/Trika-MONOREPO.git

2️⃣ Backend Setup

cd trika_backend
# Insert backend startup steps here

3️⃣ Dashboard Setup

cd trika_dashboard
# Insert dashboard startup steps here

---

🧪 Model Notebooks

notebooks/
├── workout_classification_training.ipynb
└── pose_feature_extraction.ipynb

---

🏗️ Tech Stack

Layer	Tools
Frontend	React / Next.js
Backend	FastAPI / Django
AI/ML	PyTorch, TensorFlow, MediaPipe, OpenCV
Vector / RAG	LangChain, Chroma/Pinecone
Database	MongoDB
Deployment	Docker

---

📌 Roadmap

• Mobile Companion App (Flutter/React Native)
• Live Form Correction with Voice Feedback
• Social Fitness Challenges
• Wearable Integration (Garmin, Fitbit, Apple Health, etc.)

---

🤝 Contributing

Contributions, research improvements, and dataset enhancements are welcome.

---

📧 Contact & Links

Resource	Link
Portfolio	https://kelvinportfolio2071.netlify.app/
LinkedIn	https://www.linkedin.com/in/a-anto-kelvin-prabhu-48385b25a/
Backend Repo	https://github.com/kelvinprabhu/Trika.ai_BACKEND
Dashboard Repo	https://github.com/kelvinprabhu/trikaweb_dashboard

🎯 Summary

Trika.ai blends computer vision, human biomechanics, and LLM-powered coaching to make fitness training more intelligent, personalized, and scalable.

Model Architecture