The Ghibli hype is everywhere. Our feeds are full of anime-style images – thanks to AI. And yes, it’s fun. But what often gets overlooked is the real breakthrough behind it:

With the new GPT-4o model, I’m not just mimicking cool styles – I can create exact replicas of any visual style. From any image.

My favorite source: old book illustrations

I love this website: oldbookillustrations.com A treasure trove of historical drawings – from delicate ink work to bold woodcuts. It used to be a place to browse and get inspired. Now it’s a goldmine for style references.

Not just inspired – truly style-matched

I upload an illustration, describe the scene I want – and the model merges both. The result looks like the original artist drew it today. No “AI feel”, no “kind of similar”. Just: spot on.

A design system for startups – without a big budget

For young teams, this is a real gamechanger: Creating a consistent visual style for your app, website, or pitch deck used to be expensive. Now? All it takes is a few good examples and a well-written prompt. It’s fast, unique, and looks incredibly polished.

More creative freedom. Fewer compromises

This new capability is changing how I design. No more stock images or generic templates. Now I decide how my product looks – with a visual language that fits my story.

Curious: Which old styles or forgotten artists would you bring back to life? Send me your favorite illustration – I’ll try it out for you.