
CycleStrong: Another AI Experiment!
14 Jun 2025
Repping all the Gym Girls out there
I've been lifting for 3 years now and it has completely changed my life, health and self image. I started out with a trainer to learn the ropes (SO Pawan, thank you brother), and eventually found science-based fitness Youtube. Content creators like Dr. Mike Israetel (Rennaisance Periodization), Jeff Nippard and Krissy Cela were the people I learned the most from, but by no means make up the full bredth of incredible lifting knowledge available on Youtube.
CycleStrong Coach was inspired by my love for the gym. It's essentially a cross between a gym log app and a period tracker app, combining the benefits of the two so users with uteruses can track their workouts and menstrual cycle in tandem, along with cycle phase-based prompts and tips. This time, I wanted to attempt a more complex project on Lovable. For this one, I used Supabase to experiment with actually integrating user-specific data into the app. I considered adding an AI capability, before realising that ain't free lol – maybe another time.
I learned that wiring up a database was not as hard as I thought, but it still ended up being very glitchy and tedious because of my supremely limited knowledge in this area. I took a lot more help from ChatGPT for this one generally, and that was a learning in itself- you can really make it assist you with almost anything. Well, maybe not brain surgery and the likes, but it can definitely teach me to get better at working with these back end functionalities.
Lovable really lost the plot when it came to navigating the onboarding flow when connected to Supabase user data. I spent literal hours trying to get everything to route properly, but eventually let it be since this was an experimental thing. There was probably a way to fix it, but my brain itch took over and I just decided to move on with life.
I'm now increasingly convinced that Lovable is NOT a replacement for a product designer. It's output from a design PoV is quite mediocre, not something you'd expect from a professional designer, with a lot of gap in usability, experience and overall visual impact. I ranted about it on my earlier post, if you'd like to read that. Don't forget to check out CycleStrong Coach, and tell me what you think!
PSA: The science on cycle syncing for strength is still fuzzy, and CycleStrong isn't my business idea hopping onto the influencer cycle syncing bandwagon. It's a learning exercise that presented unique opportunities WRT to user data, product design and more complex AI-assisted coding. While hormones like estrogen and progesterone can affect things like fatigue and recovery, most studies on eumenorrheic women show only minor, inconsistent differences in actual lifting performance, varying widely between individuals. Some recreational lifters feel stronger mid-cycle, others crash pre-period, and plenty don’t notice a thing. There’s no one-size-fits-all cycle centric plan to go heavy or back off. The smart move would be to track your own trends. If you feel consistently off during your luteal phase, train lighter. If you peak around ovulation, lean into it. But don’t blindly follow cycle-based plans if your body isn’t playing along. Girls, your menstrual cycle is a variable, worth paying attention to, and not a script you’re obligated to follow.