WorldClaw: Agentic 3D Open-World Generation at Scale

2026-08-05

AI Summary

Generating large-scale, freely explorable 3D open worlds demands a careful balance between global spatial consistency, rich local detail, and asset editability. To address this challenge, this work proposes WorldClaw—the first end-to-end, agent-driven framework that leverages a planning agent to parse textual inputs into structured scene specifications. This agent orchestrates the joint generation of semantic layouts, region-aware elevation fields, procedural materials, and reusable explicit assets. WorldClaw is the first method to enable complete generation of globally consistent and editable 3D worlds directly from open-ended text prompts, producing large-scale environments that are spatially coherent, visually rich, and support instance-level editing across diverse input descriptions.

Abstract

Generating large-scale, freely explorable 3D worlds from open-ended text remains challenging because a system must jointly maintain global spatial coherence, rich local content, and explicit assets suitable for downstream editing and reuse. We present WorldClaw, a fully agentic, coarse-to-fine framework for open-world 3D scene generation. Planning agents translate a text prompt into a structured specification of regions, terrain, assets, materials, and spatial relations. WorldClaw then builds a globally coherent terrain foundation from semantic layouts, reusable assets, generative or procedural materials, and a region-aware height field. For detail-demanding regions, it generates terrain-conditioned compositions, reconstructs editable textured meshes, and recovers their placement on the terrain; render-based agents further refine terrain, objects, appearance, and contacts. Across diverse open-world prompts, WorldClaw produces large-scale scenes with coherent spatial organization, visually compelling local content, and editable instance-level assets while preserving a consistent global terrain structure.

Problem

• 3D open-world generation
• spatial coherence
• editable assets
• text-to-3D
• large-scale scene

Innovation

• agentic generation
• 3D open-world
• coarse-to-fine framework
• editable assets
• terrain-conditioned composition