World Labs Launches Marble: AI Tool For Creating And Editing 3D Worlds From Text, Image, And Video Prompts
In Brief
World Labs just released Marble, its first commercial world model that generates persistent 3D environments from text, images, videos, or 3D layouts.
Spatial intelligence company World Labs has launched Marble, a pioneering generative multimodal world model that enables users to create 3D environments from text or image prompts. Initially made available to a select group of beta users, Marble now supports a broad range of input types, including text, images, video, and rough 3D layouts, and allows for interactive editing, expansion, and combination of 3D worlds. Once created, these worlds can be exported in formats like Gaussian splats, meshes, or videos, offering users more control and flexibility over their creations.
In addition to the Marble tool, World Labs has introduced Marble Labs, a creative platform that encourages collaboration and experimentation across fields like gaming, VFX, design, and robotics. Marble Labs features case studies, tutorials, and documentation, providing users with the resources to learn and share their own 3D worlds. The new tool’s multimodal capabilities are rooted in the idea that human perception of the world integrates multiple senses—sight, sound, touch, and language—into a cohesive mental model. Similarly, Marble’s world models are designed to synthesize various types of input to create comprehensive 3D worlds that evolve as new information is added.
Marble stands as a next-generation world model, offering expansive creative possibilities by allowing users to easily generate detailed 3D worlds and iteratively refine them with precision, from simple prompts to complex interactive edits.
World Labs Empowers Users To Create, Customize, And Expand 3D Worlds From Text, Image, And Video Prompts
Marble allows users to generate full 3D worlds from a single image or short text prompt, providing an intuitive and simple way to create a variety of scenes in different artistic styles. By utilizing image prompts, users can integrate Marble with other AI tools, creating 3D environments from generated images. However, while text and image prompts are powerful, they provide limited creative control, requiring Marble to generate any missing details of the scene, which can sometimes result in unexpected outcomes.
For users seeking more control, Marble also supports multi-image prompting, where different images can be combined to create a unified 3D world. This approach allows for more precision by letting users define the appearance of the world from various angles. Multi-image prompts can also be used to recreate real-world spaces, where Marble synthesizes multiple photos or videos from different perspectives into a 3D model of the actual location.
Marble’s creative process is designed to be iterative, with users able to refine their worlds further after generation. The platform includes native editing tools for making both minor and major adjustments to the 3D world, from removing or modifying objects to completely altering the visual style or structure. For more detailed control, Marble introduces Chisel, an AI-driven tool for advanced users, enabling the direct sculpting of worlds in 3D. Using simple 3D shapes or imported assets, users can structure their scenes and add text prompts to dictate the visual style, offering a new level of flexibility in world creation.
For larger and more expansive worlds, Marble provides tools to increase the size of a generated world. Users can expand specific regions, improving detail and clarity in areas that may have previously lacked precision. Additionally, Composer Mode allows users to combine multiple generated worlds into larger spaces, offering total control over the layout and arrangement of the scenes.
Once a world is created, Marble offers several export options. Gaussian splats provide the highest-fidelity representation of 3D scenes, while triangle meshes can be exported for use in various industry-standard tools. Marble also supports video rendering with pixel-perfect camera control, allowing users to capture and share their worlds in video form, with the option to enhance these videos with additional details or dynamic elements.
Marble Labs offers a platform for exploring and experimenting with world-building, showcasing creative and technical projects from artists, engineers, and designers across industries. As Marble evolves, it aims to expand into more interactive experiences, potentially enabling simulations, robotics, and other advanced applications.
Although gaming and virtual reality are often the first industries that come to mind, the potential applications for world models are vast and extend far beyond these sectors. These models have the capacity to transform various fields, including robotics simulations, architectural design, and cinematic world-building. Much like image and video generation models, world models are quickly becoming valuable tools that are expected to integrate into workflows across a wide range of industries in the near future.
Disclaimer
In line with the Trust Project guidelines, please note that the information provided on this page is not intended to be and should not be interpreted as legal, tax, investment, financial, or any other form of advice. It is important to only invest what you can afford to lose and to seek independent financial advice if you have any doubts. For further information, we suggest referring to the terms and conditions as well as the help and support pages provided by the issuer or advertiser. MetaversePost is committed to accurate, unbiased reporting, but market conditions are subject to change without notice.
About The Author
Alisa, a dedicated journalist at the MPost, specializes in cryptocurrency, zero-knowledge proofs, investments, and the expansive realm of Web3. With a keen eye for emerging trends and technologies, she delivers comprehensive coverage to inform and engage readers in the ever-evolving landscape of digital finance.
More articles
Alisa, a dedicated journalist at the MPost, specializes in cryptocurrency, zero-knowledge proofs, investments, and the expansive realm of Web3. With a keen eye for emerging trends and technologies, she delivers comprehensive coverage to inform and engage readers in the ever-evolving landscape of digital finance.
