FormCheck.AI

Helping people train safer without a trainer.

Comment

Inspiration

Most people work out without a personal trainer. They rely on guesswork to judge their form and unknowingly put themselves at risk of injury. This problem becomes even more significant for people training at home, where professional guidance is not easily accessible.

The inspiration behind FormCheck.AI was simple: What if a workout video could provide the same kind of feedback as a personal trainer? That question led to building an AI-powered system that understands human movement and provides actionable, visual corrections instantly.

What it does

FormCheck.AI analyzes workout videos to detect exercises, evaluate form, and identify mistakes with precise timestamps. Using AI-generated visual posture corrections, it shows users exactly what’s wrong and how to fix it. The app helps people train safer, improve faster, and reduce injury risk without needing a personal trainer.

How we built it

FormCheck.AI is built as a web application using React for the frontend and was vibe-coded using Google AI Studio, enabling rapid prototyping and iteration. Workout videos are processed by extracting key frames, which are sent to Google Gemini 3 Pro for temporal understanding of movement across frames. Gemini identifies the exercise, evaluates form quality, and detects issues with precise timestamps. For high-impact mistakes, Nano Banana Pro generates posture-aware visual corrections, highlighting incorrect and corrected form side by side. All analyses can be saved locally and reviewed later in a Log Book for progress tracking.

Challenges we ran into

Key challenges included iterating on prompts to generate clear and effective visual corrections, while ensuring a smooth, responsive user interface and overall performance.

Accomplishments that we're proud of

We’re proud of delivering an end-to-end AI-powered workout analysis experience with clear visual corrections and a smooth user interface. Rapid prototyping using Google AI Studio allowed us to focus on solving the core problem, while much of the surrounding infrastructure and setup was handled efficiently.

What we learned

We learned the power of Google Gemini 3 Pro and Nano Banana Pro in solving real-world problems through temporal video understanding and visual generation, and how these models can be applied across many use cases beyond fitness. Getting comfortable with Google AI Studio and Antigravity significantly accelerated our workflow, enabling rapid prototyping and iteration. This experience reinforced our confidence in building similar AI-powered tools and MVPs efficiently in the future.

What's next for FormCheck.AI

Next, FormCheck.AI can be extended to support real-time, live form analysis using the Gemini Live API. We also plan to add personalized workout planning based on available time, target muscle groups, and user goals, along with deeper progress tracking and coaching insights. These enhancements would move FormCheck.AI from post-workout analysis to a fully interactive, real-time training assistant.

Built With

Share this project:

Updates