Inspiration

  • Currently, most gym apps focus on one issue, calorie tracking or gym workout logging. We combined an app that can do both that tailors macros based on your biometric data from a 3D body scanner and personal information. In addition to this, we wanted people to use this app to make more friends at the gym. At the gym, most people have Airpods on, so it's hard to go up and talk to them. So we made an app that once you select your gym, you can see trending songs at the gym currently and if your friends are currently at the gym.

What it does

FitMax is a native iOS app structured around four main features:

Calories: Log meals effortlessly using a live camera feed or manual text entry. The app passes inputs to Gemini, returning precise breakdowns of food name, weight (grams), calories, protein, carbs, and fat. Users can review, adjust portion sizes/servings, view meal suggestions based on remaining macros, and track day-by-day history.

Dashboard: At the top, a calorie ring and macro progress bars summarize daily targets. Below, users interact with a personal virtual pet (like Biscuit the dog) that reacts dynamically to their daily progress. Logging a meal earns +10 XP, while completing a workout grants +100 XP to level up the pet.

Workout: Save custom routines, open detailed exercise views, and complete workouts to gain XP and keep active streaks going.

Your Gym (Social & Music): Select your gym from a dropdown to unlock a live Trending Songs chart for that location. Tapping any song reveals a listeners sheet showing who is listening right now or listened this week, complete with a button to open the track in Spotify. A friends grid shows who is currently at the gym. Clicking a friend opens their profile (level, streak, current song, and pet). Private profiles are supported—private members still contribute to song counts, displaying as "N more listeners have a private profile."

How we built it

Frontend: Built natively in Swift for iOS. AI Food Analysis: Integrates Gemini (gemini-3.8-flash) via REST API utilizing structured JSON schema outputs to return strongly-typed nutrition objects. Camera System: Combines an AVFoundation live camera view and image picker for quick photo analysis. 3D Body Scanning: Integrated Sam3D (an open-source model) to extract biometric data from body scans. Backend: Powered by Python and Firebase for real-time social data, user profiles, and song trending algorithms.

Challenges we ran into

Midway through development, we realized our initial web prototype didn't deliver the seamless mobile experience we envisioned for gym goers. Pivoting to a fully native iOS app meant restarting from scratch under tight hackathon time constraints.

Accomplishments that we're proud of

Successfully transforming the gym experience into a privacy-respecting social environment. We're proud of seamlessly connecting live music trends, real-time location activity, and friend interactions without invading user privacy.

What we learned

During early iterations, unstructured output from Gemini frequently caused parser errors. Implementing structured JSON outputs taught us how crucial strict schemas are when transforming large language models from conversational chatbots into reliable production backends.

What's next for FitMax

Interactive multiplayer features for pets (e.g., visiting friends' pets, joint activities). Being able to message your friends on the app Expanded pet selection, customizable accessories, and additional avatar options. Expanded integrations with popular wearable platforms (Apple HealthKit / Apple Watch).

Built With

Share this project:

Updates

Submission history