Most people come to Lovable to ship a landing page. Zohar came to build a vision.
Originally a 3D animator, Zohar Vanunu had zero coding experience—but plenty of ambition. Frustrated by the fragmented landscape of AI tools and costly subscriptions, he imagined a unified platform tailored for creatives like himself. When every “all-in-one AI platform” disappointed him, he figured: why not build his own?
The First Prompt That Changed Everything
Zohar already had a full Figma design ready when he stumbled upon Lovable on Product Hunt. Curiosity led to one prompt. Then magic happened.
I already had the design. I just wanted to generate real components for my devs. But once I saw Lovable working, I said to my wife: ‘Don’t talk to me this weekend. This is a moment in history.’ I did 1,500 prompts in two days.
That free weekend changed everything. Zohar won the internal Lovable promptathon and received 12,000 credits. He didn’t waste a single one. Welcome to MagiCan.
From Idea to Impact: Building Courses, Platforms, and Tools
Zohar used those credits to teach others. He launched the largest AI-building course in Israel—with 600+ students to date—teaching people how to go from zero to MVP with Lovable.
He also built his dream platform, Kolbo.AI (lovable version & [developers version](<>)): a mix of internal tools, Notion-like docs, and a Miro alternative—fully generated with AI, then handed off to developers for polish.
Lovable saved me at least $200K in dev costs. I did 80-90% of the work, and my devs only had to clean it up. Instead of getting stuck for two months on one feature, I iterate everything myself, visually.
Teaching the Fundamentals
In his course, Zohar emphasizes three core principles:
- Start with the basics: Blank project → Auth → Roles → Admin panel.
- Respect the backend: “If it’s not saving to Supabase, it’s not real.”
- Document everything: Context is king. He even built a Notion alternative for this.
I tell students: One prompt can make a pretty UI. But if it doesn’t work, it’s useless. Learn how Supabase works. Build secure flows first.
He also gives students a clear warning:
I see this all the time: people build a beautiful app in Lovable, but nothing is saved in the database. Don’t skip the fundamentals.
Build What You Know, Not What Sounds Cool
Zohar believes the best way to learn is by solving problems you care about:
If you’re a lawyer, marketer, or artist—build a project that solves a problem you experience. Don’t try to make something cool. You won’t stick with it, and you won’t learn from it.
That mindset helped him stay patient through debugging and iteration. And it gave him the clarity to teach others.
Debugging as a Superpower
Like many early users, Zohar hit rough patches—especially with file conflicts, API connections, and Supabase syncs. But instead of rage-quitting, he leaned in:
Every time something broke, I stopped building and started debugging. I learned how GitHub folders work. I learned how to use chat mode better. Lovable taught me how to think like a dev.
One specific insight saved him hours:
When you restore a version in Lovable, it doesn’t restore changes you made in Supabase. Once I understood that, everything changed.
Pro Tips from Zohar
- Plan prompts like a storyboard: Start with Claude or ChatGPT to outline your features, then break it down into prompt chains.
- Let AI lead, then refine: “Sometimes I just see what the AI builds, then guide it from there.”
- Leave the landing page for last: “At that point, Lovable knows your app well. The landing page is way better.”
- Embed smartly: “Use Google Drive links to embed videos. Build your own video player. Don’t pay for Vimeo.”
- Keep it simple: “Don’t build Stripe, Facebook, and WhatsApp at once. Just make the CRM work first. Get feedback. You’ll be surprised what users actually care about.”
Learning to Code? Only When Necessary
Zohar didn’t come in with a dev background. He took one Udemy course years ago, gave up after a week, and only remembered what a boolean was. Yet Lovable forced him to learn the right bits at the right time. See for yourself.
Now I can change stuff in the code. Not perfectly, but I understand what’s going on. The only place I still struggle is when I need to integrate APIs or custom edge functions. That’s when I wish I knew more dev.
Why Lovable Felt Different
Zohar has tried other platforms like Base 44 and Bolt, but found them lacking:
Base 44 doesn’t show you the code being written live, which can make you feel dumb. Lovable doesn’t do that. it treats you like you belong here and shows you its process.
It’s not about hiding code. It’s about psychology. Lovable lets me feel like I’m a real developer. That’s what makes it magic.
What's Next?
Zohar now builds Lovable-based platforms for clients and continues to explore what's possible—like his all in one AI Content Creation platform - Kolbo.AI, which all its base code made with Lovable. He plans to include all the best models in one place, offered through a single subscription that features team collaboration and project management.
He’s also asking for deeper Stripe integration: the ability to create and sync products directly from the UI, like GoHighLevel does.
It’s crazy I need to make all the products manually in Stripe. If Lovable could do that for me, it would save so much time.
He sees collaboration and workspace handoffs as the next big unlock—and he’s already testing them.
Now I don’t need to open Miro or Notion. I use the platform I built in Lovable. That’s the ultimate validation.
Lovable makes me feel like I’m not stupid. That’s the secret. You feel like a genius even if you’re not one yet.
