FormIQ AI

This is a real time ai gym coach perfect for introverts like me and for people who wants to save time going to gym and the super-high cost of gym trainers. This AI gym coach has multiple exercises.

Comment

Inspiration

Many people work out without access to a personal trainer, which often leads to incorrect form, poor posture, and higher risk of injury. Even when fitness tutorials are available online, they cannot observe the user in real time or give personalized corrections. This inspired us to build FormIQ AI, a real-time AI gym coach that uses computer vision to analyze workout posture, count reps, and provide instant feedback so users can train smarter and safer from anywhere.

What it does

FormIQ AI is an AI-powered fitness assistant that watches the user through a webcam and analyzes their exercise form in real time. It detects body posture using pose estimation, tracks movement during exercises, counts repetitions, and gives corrective feedback when the user’s form needs improvement.

The system supports multiple workout movements and provides features such as:

  • Real-time pose detection
  • Rep counting
  • Form score analysis
  • AI-based workout feedback
  • Voice coaching support
  • Performance tracking through an interactive dashboard

The goal is to make fitness guidance more accessible, personalized, and interactive without requiring expensive equipment or a physical trainer.

How we built it

We built FormIQ AI using a combination of computer vision, AI coaching, and an interactive web interface.

The frontend was developed using Streamlit, allowing users to interact with the application directly through the browser. For real-time video input, we used streamlit-webrtc, enabling webcam-based workout tracking. The pose estimation and movement analysis were implemented using OpenCV and pose detection techniques to identify key body landmarks during exercise.

For the AI coaching layer, we integrated an LLM-based feedback system using Groq, which generates short and motivating workout corrections based on detected exercise events. We also added a voice pipeline so the app can provide audio-based coaching feedback, making the experience closer to having a real trainer beside the user.

The system was structured into separate modules for video processing, exercise tracking, session state management, AI feedback, and UI rendering to keep the application organized and scalable.

Challenges we ran into

One of the biggest challenges was handling real-time video processing inside a web app while keeping the experience smooth and responsive. Since webcam frames need to be processed continuously, we had to carefully manage performance and avoid unnecessary computation.

Another challenge was integrating AI feedback with live workout events. The system needed to decide when feedback should be generated without overwhelming the user with too many messages. We also faced deployment issues related to dependencies, WebRTC support, API keys, and cloud environment configuration.

Building a voice-enabled experience was also challenging because browser permissions, microphone access, and cloud deployment behavior can differ from local testing.

Accomplishments that we're proud of

We are proud that FormIQ AI brings together computer vision, real-time interaction, and AI coaching into one complete fitness product. The project does not just display pose landmarks; it converts movement data into meaningful workout guidance.

We successfully created a system that can detect body movement, count reps, analyze form, and generate coaching feedback in real time. We are also proud of the clean product-style interface and the way the app presents fitness analytics in a simple, beginner-friendly manner.

Most importantly, we built something practical: a tool that can help users improve their workouts, reduce form mistakes, and make AI-assisted fitness more accessible.

What we learned

Through this project, we learned how to build a real-time AI application that combines multiple layers: webcam streaming, computer vision, state management, LLM-based feedback, and cloud deployment.

We gained practical experience with pose estimation, real-time frame processing, Streamlit app architecture, WebRTC integration, API-based AI services, and handling deployment issues in production-like environments. We also learned that building an AI product is not only about the model; the user experience, latency, reliability, and feedback timing are equally important.

This project helped us understand how AI can be applied beyond chatbots and used as an interactive assistant in real-world physical activities.

What's next for FormIQ AI

Next, we plan to improve FormIQ AI by supporting more exercises and adding more accurate form analysis for each workout. We also want to personalize feedback based on the user’s fitness level, workout history, and improvement over time.

Future improvements include:

  • More exercise categories such as squats, lunges, push-ups, curls, and shoulder press
  • Better form scoring using angle-based biomechanical analysis
  • Progress history and workout reports
  • Mobile-friendly interface
  • Personalized workout plans
  • Injury-risk detection
  • Multilingual voice coaching
  • Gamified achievements and streaks

Our long-term vision is to turn FormIQ AI into a complete AI fitness companion that helps people train safely, confidently, and consistently from anywhere.

Built With

Share this project:

Updates