Articles mentioning Atlas (2)
01Meta_Muse_Spark_World_Labs_Atlas_Gemini_3.8_AI_Beyond_Chat_Explained
02World_Labs_Atlas:_Single_AI_Model_Turns_Photos_Into_3D_Worlds
Key Features
Multimodal world generation: text-to-3D, image-to-3D, and video-to-3D scenes
Persistent 3D geometry that stays consistent as the camera moves (no morphing or scene drift)
Browser-based editor with camera controls, framing, and scene tweaking
Stylized and photoreal output options, viewable in the web portal and VR headsets
Export of generated worlds (Gaussian splats / mesh formats) into Blender, Unity, and Unreal pipelines
Pros & Cons
Pros
Built by a top-tier research team (Fei-Fei Li, Justin Johnson, Christoph Lassner, Ben Mildenhall) with deep 3D/NeRF expertise
Geometry is genuinely persistent and 3D-consistent, a key weakness of the video-generation-based world models
Free experimentation tier and an approachable web interface lower the barrier for non-3D specialists
Cons
The 'Atlas' name could not be verified against World Labs' official channels, so product details and roadmap are unconfirmed
Access has been staged via preview/waitlist rather than open signup, and generations are compute-heavy
Outputs often look splat-like or low-fidelity and typically need cleanup in a real 3D pipeline
No widely available public API, mobile app, or documented enterprise/on-prem deployment