Can We Create a Cat-Food Subscription Service with vibe coding?

There's a lot of buzz around vibe coding these days. So, we thought, why not put some of the top ones to the test? We chose four popular platforms—Lovable, TempoLabs, Replit, and Bolt.new—and challenged them with a fun yet practical task: building an online shop for a healthy cat-food subscription service. After all, who doesn’t want to make sure their feline friends get their organic, vet-approved meals right on time? 😉

The Challenge:

We wanted to see if these vibe coding tools could go beyond the usual suspects like landing pages or basic to-do apps. Could they handle something more complex, like a full-fledged e-commerce site with subscription capabilities?

We decided to start with four tools and see how far we could get without wasting ages fixing errors. Meaning: if we hit a wall with one tool, We’d switch to the next one. Simple timeboxing. After all, these tools all promise quick results. So, let’s see who keeps their word! 😉

The Prompt:

To keep things fair and square, we gave each platform the same instructions:

"Create an online shop for a healthy cat-food subscription service. Cat owners can buy healthy, organic cat food with a subscription service. Main value propositions: Convenience: never run out of cat food with automated deliveries. Customization: Meal plans tailored to a cat's age, health, and preferences. Health & Quality: Vet-approved organic, high-quality ingredients. Flexibility: Ease subscription management (pause, change, or cancel anytime). Create the online shop from homepage to checkout and account management."

General Observations:

We kicked things off with all four platforms. Interestingly, they all initially focus on generating the frontend designs and workflows—probably to show something shiny fast. Things like data storage, validation, error handling, and so on? None of them touch that part at first.

Funnily, each one came up with exactly the same product name 'PurrFect Meals' 😸, but it also makes sense as they're all using the same LLM. Without providing specific direction for unique branding, it naturally comes up with the most obvious cat-food wordplay, which apparently is 'PurrFect Meals.

‍

Bolt.new

The generated frontend is pretty basic, but functional. What felt a bit off though, was how they mixed up the product setup (the cat food recipes) with the subscription plan itself. We’d have preferred a clear distinction—either proper subscription packages or product bundles—not a mashup of both. Speaking of mashups: Bolt.new uses Unsplash for images, and apparently, when it saw the word "Mix," it slapped a DJ image on the product "Garden Fresh Tuna Mix." Not exactly what we had in mind.

After the first draft, we kept running into timeouts, API overloads—you name it. Frustration levels definitely climbed. That said, after some back-and-forth, Bolt.new powered through, and integrating Stripe as well as Supabase for data storage was possible. Integration steps came with clear instructions and extra docs in case you got stuck.

Would we have made it to the finish line? Probably. But honestly, the constant hiccups and the time it took to fix everything felt like too much hassle. So we moved on to the next tool.

‍

Replit.com

One cool feature here: Replit suggests new features you might want to implement. Handy if you’re stuck in ideation mode. But, the frontend design from the initial draft? Again, pretty basic. Plus, the flow felt odd—starting off with a login screen right away. Image choices? Zero connection to cat food. And since we'd already need to rebuild parts of the flow, I decided to call it quits on Replit too.

Side note: We’ve had solid results with Replit on other projects. It offers a lot of flexibility, especially for more experienced users or developers (authentication, deployment, databases).

Tempo

Tech stack: Vite, Tailwind, React. Tempo kicks things off by generating a Mermaid diagram of the user flows. You can’t edit the diagram, but it’s still a handy guide. First impressions? The frontend is definitely the best-looking one of all the tools—colors and image choices are spot on, and the overall layout feels pretty polished.

However, Tempo stumbled out of the gate, throwing up a few errors. Luckily, their “Fix it” button sorted those out. But when we tried to integrate Supabase, more errors popped up. Tempo claimed they were fixed, but suddenly empty screens appeared in random places. Shame, because the initial setup was really promising. They even added a cat assessment quiz as the first step—nice touch! Supabase and Figma were already integrated, and their visual editor made frontend tweaks super simple.

But, prompting speed was noticeably slower compared to Lovable. Combined with long wait times and recurring errors, Tempo didn’t make the cut for me this time.

Lovable.dev:

Last one up—and the only European player in the mix. First draft looked solid, though we couldn’t quite figure out why the main image was a cat painted in Renaissance oil painting style. 😂 But hey, at least it stood out from the others.

One thing we loved: Lovable pulls inspiration references during initial setup (cool!), and instead of creating bundles, it generated multiple individual products—which matched our original idea better.

First draft done, next step: Supabase integration. And it worked flawlessly. In no time, we had user management switched over to Supabase, a login/signup page, and the product suggestions stored in a database and dynamically loaded (one prompt was enough here!). So far, smooth sailing.

Tried adding Stripe integration—after a few “Fix it” iterations, it was live! 👋🏻

Without any serious coding knowledge, we had a working online shop with user management, product management, and a Stripe checkout up and running in no time. Sure, plenty of iterations would still be needed for error handling, subscription management, and UX polish—but overall, it works! Yeah, we know, you could also do all this with Shopify. 😅 BUT, we are super happy with the progress we were able to make in a few hours spent with prompting.

Check out the current result here: Cat-food subscription service with Lovable

‍

‍

Conclusion

Lovable and TempoLabs stood out with the most appealing results. Their frontends looked more polished and professional than the others. With Lovable, we had an online shop with account management, dynamic product database, and a working Stripe checkout running in no time. Crazy stuff. On another day, maybe with another use case, we might’ve gotten similar results with the other tools. So, all four definitely have their place.

What’s truly wild is how far you can get in such a short amount of time with these tools. According to Anthropic’s CEO (the folks behind Claude Sonnet 3.7), coding quality is expected to improve massively this year alone. Check out the article here.

That said, we’d love more control—being able to tweak the code directly or having better frontend management through a proper visual editor instead of relying solely on prompts. Prompt-first approaches aren’t inherently bad, but having to do everything through prompts feels clunky more than helpful.

Also, big topic (though out of scope for now): Quality blind spots and security vulnerabilities. Yep, those are issues too.

‍

Our Takeaway

So far, we've used these systems mainly for rapid prototyping. But development is moving fast, and with enough iterations, fully functional products are definitely within reach. What’ll be exciting to watch is how established no-code platforms like Bubble and Flutterflow—both backed by huge communities—keep up. We wouldn’t count them out just yet.

In conclusion, there’s no one-size-fits-all solution. Each tool has its own strengths, and more players will surely pop up.

If you want to see how we rate other products, check out NAUT.

We’ll definitely keep testing and share more soon! 😎