Architecting the Spatial Web: A Review of the Sneeze Metaverse Browser Engine and Multi-Origin Scene Composition

Abstract

This technical report examines the architectural case for a dedicated Metaverse Browser Engine (MBE), as put forward by the Open Metaverse Browser Initiative (OMBI), an open-source, vendor-neutral testbed hosted under the Metaverse Standards Forum (MSF). The report is not original empirical research; it is a structured synthesis and critical discussion of publicly disclosed architectural material presented by MSF/OMBI and its RP1 prototyping effort at a W3C Immersive Web Community Group discussion held in New York in May 2026, released under a Creative Commons Attribution 4.0 International license. The purpose of this report is to make that material accessible to a broader research audience and to situate it within the wider immersive-web and spatial-computing standardization landscape. The report first identifies why current web and WebXR browser stacks lack a browser-level model for multi-origin spatial scene composition, automatic geolocated service discovery, persistent shared geospatial anchoring, and a standard live 3D service abstraction. It then describes the proposed architecture, centered on Sneeze, an open-source Metaverse Browser Engine intended to serve the metaverse stack in a role comparable to that of WebKit or Blink for the two-dimensional web. Sneeze is built around a Scene Object Model (SOM) as its central scene graph, a Remote Model Access Protocol (RMAP) for synchronizing service state with clients, WebAssembly-sandboxed per-service execution, and a certificate-based, organization-scoped trust model for multi-origin content. Four architectural patterns for composing spatial (Sneeze) and conventional web (Blink) rendering are reviewed, together with open standardization gaps in geospatial pose and anchoring interoperability. The report concludes that a dedicated, complementary spatial browser engine is a technically motivated and plausible path toward an open, decentralized spatial web, while noting that the architecture remains at an early, testbed stage of maturity.