CloudFitness

CloudFitnes React
Architecture Diagram

Inspiration

Getting started is always the hardest part, and I was no exception. I was lucky—my cousins had already been through it and took the time to guide me through those uncertain first steps. They showed me what to do, how to do it, and made the beginning feel far less overwhelming.

That experience made me realize that many others probably feel the same way: lost, unsure, and wishing they had someone to guide them. And that’s what inspired the idea—if we could create an app that offers that kind of direction and support, we might help more people get into fitness with confidence instead of confusion.

What it does

CloudFitness Overview

CloudFitness is your AI-powered fitness partner, built directly into your browser. It guides you through every stage of your journey with four core tools.

AI Form Coach Upload an exercise video and get a clear, frame-by-frame analysis powered by Gemini. It identifies the single most important fix you need and gives you a simple correction plan with explanations and encouragement.

Personalized Starter Plan Answer a few quick questions, and the AI generates a customized 4-week plan based on your goals, body, and lifestyle—including a workout split and basic nutrition guide.

Interactive Progress Book Log daily photos, weight, and notes. Your entries are transformed into a dynamic flip-book that visually captures your transformation.

Activity Dashboard Your homepage features a GitHub-style consistency graph and quick access to all tools, keeping your progress easy to track and understand.

How we built it

CloudFitness is built on a modern, privacy-focused tech stack designed for speed and reliability. The frontend uses React and TypeScript for a clean, maintainable interface, with Tailwind CSS providing a responsive, dark-mode-ready design. At the core is the Google Gemini API: we send video frames for multimodal form analysis and rely on structured JSON outputs to keep responses consistent and easy to integrate. All user data stays on the device using IndexedDB through the idb library, ensuring privacy and offline access. To elevate the experience, we built custom UI features like a smooth video trimmer, animated report visuals, and a page-flip progress book that makes tracking progress feel fun and intuitive.

Challenges we ran into

Refining the AI Coach’s eye was a major challenge. It took many prompt iterations to get Gemini to behave like a nuanced kinesiologist—recognizing that “perfect” form varies by person—while still delivering consistent, structured JSON feedback.

Accomplishments that we're proud of

We’re proud of the depth and quality of the AI’s analysis. It goes beyond spotting errors, delivering constructive, safe, and encouraging feedback that feels like it’s coming from a real coach.

What we learned

Gemini’s structured JSON output is a game-changer. Being able to force responses into a predefined schema makes AI apps far more reliable—and saved us countless hours of development and debugging.

What's next for CloudFitness

We plan to introduce real-time feedback using the Gemini Live API so users can receive instant, audio-based form corrections through their device’s camera. Alongside that, we’re expanding the exercise library to cover more movements—everything from yoga and HIIT to advanced weightlifting—so the AI can coach a broader range of workouts with accuracy.