In October of 2024
I began building a weightlifting web application. I worked on the web app until transfering to an iOS app on May 22 2025. On Aug 1, 2025 I launched Hybra to the app store. Both applications are supported by similar APIs built with Typescript, Node.js, Hono, and Cloudflare Workers. The idea began as a fun project to support my hobby of learning through building. While I have enjoyed the process of building this app, the project is fighting for real estate in my mental bandwidth and I want to be developing something that people find useful.
Today it is August 27
and my goal is to record what I have learned and potentially extract some lessons for future endeavors.
Building something you love is easy — getting others to love it is a challenge.
Lessons
- Users are hard to acquire.
- I would start with idea and market first instead of building
- Build analytics from the ground up and use them to make decisions
- Figuring out when to switch ideas is a tough decision
- Cloudflare workers are a great cheap option for backend deployments
- Cloudflare pages was clunky developing locally due to slow build times
- I like deploying web apps with Vercel over Cloudflare
- Swift types are checked at compile time. This slowed my development while learning as there was less room for error and resulted in cryptic compile errors that resulted in having to search through the code.
- AI coding tools are really fun when making MVPs but their results often begin to get worse and worse the code complexity increases. Fast times turn into bugs and AI rewriting code for no reason.
- Setting up domains in 2025 is a breeze.
- Using one google account to register domain and users is a functional strategy. It allows for envolving domain name over time and maintaining accounts. Downside is $26 monthly.
💰 Cost Breakdown (as of Aug 2025)
| Service | Cost | Notes |
|---|---|---|
| Cloudflare Workers | $5/month | Budget-friendly serverless backend |
| SendGrid (email) | Usage-based | Cost depends on volume of emails |
| iOS Developer Program | 8.33/mo) | Required for App Store distribution |
| OpenAI API | Usage-based, “cheap” | Less <$20/month at current scale |
| Domains (2) | 1.67/mo) | Standard registration |
| Google Workspace | $26/month | Domain + accounts |
Estimated recurring monthly cost:
- Fixed: **5 Cloudflare + 8.33 Apple + $1.67 domains)
- Variable: SendGrid + OpenAI (likely < $20 combined at current usage)
So, I’m sitting around $45–60/month total burn right now.
I am the only user
At this point, I am the only active user of Hybra. Around 20 people have signed up — half of them friends and family — but no one has returned consistently. This is an important reality check.
What this really means is that while I succeeded in building and launching a real product across web and iOS, I never validated whether the problem I was solving was compelling for others. Hybra solved my itch as a weightlifter, but I never proved that others shared the same pain strongly enough to come back.
This is valuable data. It’s a reminder that building without validating demand first can lead to beautiful products that no one adopts. For future projects, I plan to start with the market: talk to potential users, validate their needs, and then build only what addresses the clearest pain points. That way, traction can be tested before investing months of coding effort.
🔮 Extracted Lessons for Future Projects
- Validate with users early. Don’t build a full stack until you’ve tested market interest.
- Cheap + fast iteration wins. Vercel and Cloudflare are great for MVP speed.
- Budget realistically. Even “cheap” side projects creep up toward $50+/month.
- Balance AI use. Great for boilerplate, dangerous for complex refactors.
- Set a review checkpoint. Decide in advance when you’ll stop, pivot, or scale.
Cheers!