LITTLE SPAIN × ZARA

LITTLE SPAIN × ZARA

2025

Little Spain / Zara

Hybrid VFX / Brand Campaign

Overview

Madrid covered in snow. A logistically impossible brief — solved through a hybrid model combining real location footage with AI-generated VFX. Cinematic result, fraction of the cost, fraction of the time.

Concept

The brief was simple and impossible at the same time: put Zara in a snow-covered Madrid. No waiting for weather. No location move. No CGI budget at traditional scale. The answer was not to fake it — it was to build it in layers. Real city, real light, real camera movement. AI-generated snow integrated with the coherence of something that was always there.

Notes

A collaboration between Little Spain (production, location filming) and MITO (AI-native VFX and finishing). This project represents the hybrid model at its most direct: neither party could have delivered this alone at this quality and cost.

LITTLE SPAIN × ZARA — overview

Minimal inputs

This project began with almost no traditional production material: just two phone photos of the characters, already wearing the exact wardrobe that needed to appear in the final piece. From those images, combined with Google references and environment research, we built every shot of the project.

The challenge was not simply to generate new visuals, but to expand a minimal starting point into a fully coherent visual sequence. We used those base images to preserve the identity of the characters, maintain wardrobe accuracy, and construct scenes that felt spatially, atmospherically, and cinematically consistent across the entire piece.

From reference to sequence

From there, we developed the full visual system: framing, environments, lighting, continuity, texture, mood, and integration between subject and setting. The goal was to make each image feel like part of the same production, not like a collection of disconnected generated outputs.

A new production model

That is where this model becomes powerful. It is not about creating images from nothing. It is about using limited but meaningful real inputs to create visual systems that are scalable, believable, and precise enough for brand storytelling.