Marble by World Labs

Product Introduction

1. Marble by World Labs is a generative AI platform that enables users to create high-fidelity, persistent 3D worlds using multimodal inputs such as images, videos, text prompts, or 3D layouts. It leverages World Labs’ proprietary multimodal world model to automate complex 3D design processes, making advanced world-building accessible to non-experts. The generated environments are fully editable and interoperable, supporting applications in gaming, virtual events, architectural visualization, and immersive storytelling.
2. The core value of Marble lies in democratizing 3D content creation by eliminating the need for specialized skills or manual modeling workflows. It reduces the time and cost of building detailed 3D environments while ensuring consistency and scalability across projects. By integrating persistence, Marble ensures that worlds remain editable and reusable, enabling iterative design and collaborative workflows.

Main Features

1. Marble allows users to generate 3D worlds from diverse inputs, including text prompts, single images, videos, or basic 3D layouts, using its multimodal AI engine. The system automatically interprets spatial relationships, materials, and lighting to produce optimized, ready-to-use environments. Outputs are compatible with standard 3D software and game engines like Unity and Unreal Engine.
2. Users can explore a curated library of community-generated worlds spanning categories like Stylized, Realism, Interior, Exterior, Fantasy, and Sci-Fi. Each world is tagged with metadata such as creator credits, generation parameters, and usage rights, enabling seamless remixing or adaptation. The platform supports real-time collaboration, allowing multiple users to edit or expand shared environments.
3. The Studio module provides advanced tools for editing geometry, applying textures, stitching multiple worlds into larger landscapes, and rendering cinematic sequences. Users can export worlds as GLB, OBJ, or FBX files or publish them directly to Marble’s cloud for cross-platform access. A physics engine integration enables dynamic object interactions and environmental simulations.

Problems Solved

1. Marble addresses the steep learning curve and time-intensive nature of traditional 3D modeling software, which requires expertise in tools like Blender or Maya. By automating scene generation and optimization, it enables rapid prototyping and iteration for professionals and hobbyists alike. The platform also solves fragmentation in collaborative workflows by centralizing asset storage and version control.
2. The primary target users include game developers, architects, virtual event planners, and digital artists seeking to accelerate environment design. Educators and researchers can use Marble for immersive simulations, while brands leverage it for creating virtual showrooms or marketing experiences.
3. Typical use cases include generating a sci-fi game level from a concept sketch, converting a real estate video into a navigable 3D model, or stitching fantasy landscapes for a VR film. Engineers use Marble to visualize industrial layouts, while educators build historical recreations from text descriptions.

Unique Advantages

1. Unlike single-mode AI tools like MidJourney or DALL-E, Marble’s multimodal engine interprets hybrid inputs (e.g., image + text) to resolve ambiguities in spatial design. It outputs fully structured 3D scenes with layered materials and lighting, not just 2D renders. Competitors like Promethean AI focus solely on game assets, whereas Marble supports cross-industry applications.
2. The platform introduces persistent worlds that retain edit history and parametric data, enabling non-destructive modifications. Its stitching tool uses topology-aware algorithms to merge disparate environments without manual retopology. The AI also auto-generates LOD (Level of Detail) models for performance optimization.
3. Competitive advantages include proprietary training data from partnerships with 3D asset libraries, ensuring legally compliant outputs. Marble’s cloud backend scales to handle terabyte-scale projects with sub-50ms latency for real-time edits. Enterprise-tier users gain API access for integration with BIM software and custom AI fine-tuning.

Frequently Asked Questions (FAQ)

1. What input formats does Marble support for world generation? Marble accepts JPEG/PNG images, MP4/MOV videos (up to 60 seconds), text prompts in English, and 3D layouts in OBJ or FBX formats. Video inputs are processed at 30 FPS to extract temporal spatial data, while images are analyzed for depth maps and material properties.
2. How does “persistent” 3D world functionality work? All generated worlds are stored in Marble’s cloud with version control, allowing users to revisit and modify any element (e.g., terrain, lighting) indefinitely. Changes propagate non-destructively, preserving original assets for rollback. Persistence includes metadata like AI generation parameters for reproducibility.
3. Can multiple users collaborate on a single world in real time? Yes, Marble Studio supports concurrent editing with granular permission controls. Changes are synced via operational transformation to prevent conflicts, and a history slider lets users review or revert edits. Voice/chat features are integrated for team communication.
4. What export formats are available for generated worlds? Worlds can be exported as GLB (glTF 2.0), OBJ, FBX, or USDZ files, with optional texture atlasing and animation baking. For developers, direct publishing to Unity (via UPM package) or Unreal Engine (Datasmith plugin) is supported.
5. Are there hardware requirements for running Marble? Marble’s web editor works on any modern browser with WebGL 2.0, but GPU-accelerated features require a compatible graphics card (NVIDIA GTX 1060 or higher recommended). Cloud rendering options are available for low-spec devices, with pricing based on compute minutes.