Hi Reader,

Welcome to Hal9 Weekly — insights from our founder Javier on turning AI ideas into real products.

Javier's Insights:

Lovable Is Great. Until It Isn't.

We use Lovable at Hal9 all the time for building initial demos. It's genuinely one of the best tools out there for getting from zero to something fast.

But there's a conversation founders need to have earlier than most of them do: what happens when you move beyond the prototype?

We recently ran a full experiment — built a real community app, did realistic load testing, and compared the infrastructure costs at scale. What we found was worth sharing.

Even at small scale (184 light user sessions in our test), Lovable Cloud's costs became difficult to track and optimize. There are no real-time dashboards, limited visibility into what's actually driving your database spend, and a 24-hour billing delay that makes it hard to catch problems before they compound. There's also a infrastructure markup that's reasonable when you're just starting out — but becomes very meaningful once you're scaling.

To be clear: this isn't a criticism of what Lovable has built. They're at $500M ARR and adding a million new projects a week. That's an extraordinary product. But at that scale, there's a real opportunity — and need — for better cost transparency, security, and reliability tooling for production apps.

For founders building something real, the practical move is to plan your infrastructure migration before you need it. In our experiment, moving to Supabase and Vercel (Supabase is literally the backend powering Lovable Cloud) gave us full cost transparency and a flat ~$10/month for the micro instance — capable of handling millions of user sessions reliably.

Lovable excels at speed. Production apps need thoughtful infrastructure decisions.

Build fast. But plan for scale from day one — or you'll face painful surprises later.

Full breakdown with tests, numbers, migration steps, and screenshots in the link above — worth reading if you're building anything production-grade.

Customer Success Story: DesignQA

Merging Human Creativity With AI Speed — A Website Generator in Four Weeks

DesignQA is tackling one of the oldest tensions in web design: the tradeoff between quality and speed. Their vision is to combine designer-crafted templates with AI automation, so businesses can get polished, professional websites without sacrificing the creative quality that templates alone can't deliver.

Founder Mahdi Farra knew the concept was strong. He needed to know if it was buildable — fast enough to validate with real users and map a credible product roadmap.

The Challenge: Build a functional AI-powered website generator MVP in just four weeks — one that supported template-based design, AI-generated imagery, and live website previews, while also surfacing insights into technology gaps and market opportunities to guide the next phase of development.

Hal9 helped DesignQA:

🚀 Deliver a working AI-powered website generator MVP in four weeks

🎨 Build a flexible template system using Tailwind CSS and HTML, enabling designer-driven layouts with real aesthetic control

🖼️ Integrate AI-powered image generation to create custom visuals that enhance website quality automatically

👁️ Enable static website generation for fast, reliable previews — so users can review and iterate on designs before publishing

🗺️ Identify technology gaps and market opportunities through the build process, giving DesignQA a clear, evidence-based product roadmap

The result? DesignQA emerged with a validated concept, early user engagement, and a sharper view of where to take the product next — all in under a month.

Let’s Talk

Got an AI idea/building something and want to chat about it?

Book a free consultation with Javier — he’ll help you validate it before you build, so you don’t lose time chasing the wrong version.

Talk soon,

—The Hal9 Team