The app you're using to read this blog post was created with the Replit Agent. I’d been hearing a lot about vibe coding—especially in the context of Replit—and decided to try building something for myself. I started out with an ambitious idea: a social network for software engineers. But soon after launching the first version, I pivoted to making it a personal blog with a social media feed feel. Managing multiple profiles and content creators felt overwhelming and, to be honest, beyond my skill level.

Prompting an AI to generate a fully functioning application feels almost magical. However, certain aspects make it clear that vibe coding isn’t easy, and the agent can’t be fully trusted to create production-ready code on its own. An earlier version of this app included a Stripe integration for the Sponsor Me page. I ultimately scrapped the entire integration because I didn’t have the knowledge to secure it properly. I know that exposing secrets in the console and URL was a major oversight. What’s worse, the app was live when that happened, even after I had the agent audit it for security.

Making changes can sometimes feel like playing whack-a-mole with bugs: you fix one issue, and another pops up. You have to be precise in how you prompt the agent, because when it takes a wrong turn, it spends time and credits to correct itself. That leads me to my last big takeaway from building my first vibe-coded app: it costs money to develop this way. I don’t see many people talking about that.

I burned through my monthly $25 credits in just a few hours, and I’m now at around $75 in total spend. I think I’m done tinkering with it, and I’ve resolved the security issues, but there’s still a cost to keeping the app deployed and running. Overall, though, I don’t think the expense is unreasonable. I don’t know how much it would cost to hire a developer or how many hours it would cost to do it by myself, but I’m pretty sure $75 is the cheapest of any of those options.

So, how do I feel about vibe coding after shipping my first app? I think I’m addicted! I’m already brainstorming ideas for my next project, and I plan to create a full portfolio of apps. There’s definitely a learning curve, but next time will be much smoother.