I always wanted to build a health and fitness app, That was the whole reason little me wanted to become an app developer. I came across calisthenics in my late teens and getting into it was one of the best things that happened to me. The cool calisthenics skills always left me in complete shock.

When I started, there weren’t many calisthenics apps out there, so I used general workout apps like Strong and Hevy. Logging sets was easy. Real form feedback and clear progression were not.

In my early 20s, after I’d grasped Android development, I built a personal project on pose detection and exercise form coaching. That led to work on a medical workplace posture correction app for Android and iOS, where I learned a lot about biomechanics and movement analysis.

Feeling stuck creatively, I was scrolling Reddit when I came across Shipaton. I read through the contest details, and suddenly I just knew what I had to build. Taking inspiration from anime and calisthenics, I built Rate My Form. It’s simple: record your exercise form, let on-device ML analyze the movements, and generate tips and coaching cues to help you fix inaccuracies and improve your form and reps. I also put the biomechanical knowledge I learned on that work app to use, and based on it, I developed algorithms to detect whether a user’s muscles are in tension zones and how much they actually worked the muscle.

What Rate My Form Does

Rate My Form is a calisthenics form coaching app. You pick an exercise, place your phone on a stable surface, and record your set. On-device ML analyzes the movement in real time with live rep feedback, joint tension updates, and full ROM detection. When the set is done, you get coaching cues, improvement tips, and performance metrics you can actually act on.

You can also create or join challenges with friends to sync sets and stay accountable. Form analysis works without an account. Challenges need sign-in.

Ranked turns quality training into progression. You earn XP from clean sets based on full ROM and good TUT, up to 10 XP per set, and climb from Unranked to E, D, C, B, A, S, and National. Completing a set alone does not grant XP. About 10 quality sets is roughly 100 XP. The endgame is reaching S or National and maintaining that rank.

Building Rate My Form

The building process involved getting feedback from users who tried the app, as well as communities on Reddit that shared insightful tips and helped me improve it further.

One of the steps was competitor research, gathering similar apps, checking how they function, and looking at their marketing channels. This helped a lot and gave me a clearer idea of where my users would be based and whether localization for those regions made sense.

Development of the app involved a tech stack with MediaPipe for on-device pose recognition and AVFoundation in Swift for processing frames and rendering them on screen along with the coordinates. I also built a custom pipeline to recognize inaccuracies in the exercises and provide users with voice output, metrics, tips, and coaching cues related to the workout after each set was analyzed.

The last piece of the puzzle was distribution, and it’s exciting as well as exhausting. I plan to spend more time on this, improve further, and build a brand identity around the app instead of just posting app usage videos. So far, paid advertising has brought in a good number of users, while organic is still a work in progress.

One thing I did during the building process was make the release notes worth reading. Instead of just listing technical updates and boring the user, I tried to keep them fun and light. This was inspired by Slack’s release notes. I still read their notes from time to time because they feel like a standard developers should maintain. I’ve attached my release notes, Hope you’ll enjoy reading them.

Challenges That Made Me & The App Better

During real workouts, I ran into a weird problem: the app would crash occasionally and I couldn’t pull logs at the moment it happened. To catch them, I had to train with my laptop and phone connected for a while. Eventually I got a log and it pointed to a memory leak in the thread where the MediaPipe pipeline was outputting coordinates to the UI layer. Once I found it, fixing it with AI’s help was much easier. That challenge ended up improving the app further and it gave me a deeper understanding of thread handling and background processing on iOS.

Having touched distribution, I can say I’m better now than I was before Shipaton and I don’t plan to stop after it. I plan to keep improving. Distribution didn’t just push me to polish the app frequently, it also pushed me to create good creatives and test them against ad metrics. Along the way I also made some simple 2D animations for the app’s character and it was wonderful. I plan to keep improving the brand identity, make the app more approachable for users and help them upgrade their calisthenics workouts.

The most important thing I realized along the way is that coding - bringing what you think into existence is easier than ever, with AI. Since people mainly want end results, the better combo is not relying on AI for everything but exercising your own intellect and using AI to improve and boost it. More AI-assisted coding, less slop coding.

Things I Learned Along The Way

After the app was built, it was time for distribution, a territory I had never experienced. I thought it would be easy. It wasn’t.

The Reddit Experience

I tried posting on Reddit for a week. The outcome: very few people gave useful suggestions, and some just bashed the app. For a moment I felt like, “What’s the point?” But that wasn’t the goal. The goal of Rate My Form was simple: make calisthenics fun, build a community around it, and above all help people progress toward better strength.

After gathering the good and the bad feedback, I went back to work. I improved the posture detection algorithms, personalized the coaching cues further and added support for more exercises.

The Instagram Experience

After testing organic posts on Reddit, I tried Instagram too. Since I knew video editing, I started making app usage reels, posting them and trying different creative styles to see what worked best. What I noticed with Instagram is that it can give organic reach and users but you have to play the long game, unless you’re ready to run Meta ads.

The Meta Ads & Apple Ads Experience

Along the way I also tried Apple Search Ads and Meta ads, which brought a solid number of users. I ran Meta ads on a short budget and it felt much quicker compared to organic posting. I tested different angles with the creatives as well and the ones that won gave me good reach and users.

Having just enrolled in the Apple Developer Program, I learned that Apple gives new developers a one-time $100 ads credit. I didn’t know about it at first, but once I did, I used it to boost visibility on the App Store. Most of my users came from two countries, Brazil and Spain, so I added localization for them.

The main thing I learned from this experience is that coding an app is no longer the hard part. Getting users and making them feel part of a community is. For that, developers have to play the long game of creative marketing or build a brand and personality around a niche well before the app launches.

Accomplishments I’m Proud Of

Shipping Rate My Form during Shipaton is the big one. What started as a personal calisthenics idea is now live on the App Store with on-device form analysis, coaching cues, Challenges, ranked progression, and a Pro subscription through RevenueCat.

I also learned distribution the hard way. Reddit feedback, Instagram reels, Meta ads, and Apple Search Ads all taught me something. Paid ads brought users faster. Organic is still a grind. Both pushed me to polish the app more often and make creatives worth testing.

And the small stuff counts. Fun release notes. A simple 2D character. Fixing a mid-workout crash caused by a MediaPipe memory leak. Those details made the app feel more like a product and less like a prototype.

What’s Next For Rate My Form?

Since the base structure of the app is done, I now plan to improve what’s already there and add more features later based on market research and user feedback. For now, the focus is:

1. Accountability and Notify Friend
   Improve the accountability feature and add a Notify Friend option, so a friend in the same challenge can nudge you if you haven’t done your day’s workout. Custom notifications will come too, but a personal push from a friend hits different.
2. Progress charts
   A progress chart that shows the journey across the ranks you’ve completed and how much your form has improved compared to when you started.
3. More robust live coaching cues
   Stronger live coaching so issues like shoulder inaccuracies can be caught during the set and flagged to the user in real time.
4. Brand identity, advertising and distribution
   Create a clearer brand around Rate My Form instead of just posting app usage videos. Keep testing creatives, improve organic content on Instagram and treat paid ads as something to refine, not just turn on.

Rate My Form was built with SwiftUI, AVFoundation, MediaPipe, TensorFlow, Google ML Kit, RevenueCat.