https://youtu.be/4dIVjG11jgQ

Project Overview: Graswald

Democratizing High-End Digital Nature

Graswald is a world-class 3D ecosystem designed to help artists create hyper-realistic natural environments in Blender. With a vast library of high-fidelity assets and the powerful Gscatter tool, it removes the technical barriers to world-building, allowing creators to focus on artistry rather than geometry nodes.

As the Product Designer, I was tasked with evolving Graswald from a collection of assets into a cohesive, user-centric platform. I led the design of the G-Library and asset management workflows, focusing on making professional-grade vegetation accessible to everyone—from solo hobbyists to architectural visualization studios. My work bridged the gap between complex 3D data and a seamless "click-and-scatter" interface.

Assets.png

The Problem: The "Barrier to Realism"

Creating realistic nature in 3D is notoriously difficult and time-consuming. Through user research and community feedback, I identified three primary friction points for artists:

  1. Technical Overload: Building believable ecosystems required a deep understanding of complex scattering systems and "LOD" (Level of Detail) management, which interrupted the creative flow.
  2. The Repetition Trap: High-quality assets often felt "stiff" or repetitive when scattered, breaking the illusion of nature.
  3. Performance Anxiety: Managing hundreds of high-resolution plants often led to viewport lag and system crashes, forcing artists to choose between quality and performance.

Users weren't looking for more assets; they were looking for a smarter way to manage them. They wanted a system that understood the "rules of nature" so they didn't have to.

Assets viewport.png

Ecotope display state.png

The Solution: Intelligent Ecosystems

I redesigned the Graswald experience to act as a bridge between the artist’s vision and Blender’s technical power. The solution shifted the focus from individual plants to Contextual Ecotopes.

Instead of browsing through hundreds of disconnected models, I introduced a system where assets are grouped into biologically accurate "Ecotopes" (e.g., Urban Wilderness or Garden Glory). This allowed users to: