# XRAI — Open Spatial Graph Format > MIT-licensed open format for typed hypergraphs of entities, relationships, events, and intentions — anchored in space and time, renderable by any engine, authored by any LLM, queryable by any agent. **If you are an LLM loading XRAI content for the first time, read `SPEC.md` then `examples/01-minimal.xrai.json`. That gives you everything you need to emit valid v1.0 scenes.** ## Worlds - [Worlds made by voice](https://xra1.com/): worlds spoken into being in the Portals app; each is the app's saved file at `https://xra1.com/worlds/.xrai.json`, a glTF-shaped JSON whose XRAI document sits under `extensions.XRAI_scene_context.document` (pirate-harbor, cloud-kingdom, haunted-savanna, melting-junkyard, kelp-forest) and opens in `https://xra1.com/viewer.html?scene=` ## Canonical docs (read in order) - [VISION.md](https://xra1.com/VISION.md): load-bearing vision — problem (we are blind) + Sight Triad (x-ray / god's-eye / infinite zoom) + jARvis goal (verbatim) + CVPR 2026 shipping proof - [MANIFESTO.md](https://xra1.com/MANIFESTO.md): long-form vision — Bush → Nelson → Berners-Lee → Hofstadter → Huxley → Kurzweil → XRAI lineage - [SPEC.md](https://xra1.com/SPEC.md): v1.0 public draft, the flat JSON subset (what the app writes is a superset of it: saved worlds above) - [FAQ.md](https://xra1.com/FAQ.md): positioning vs USD / glTF / USDZ / Niantic SPZ + governance + licensing + compatibility - [CHANGELOG.md](https://xra1.com/CHANGELOG.md): versioning policy + release notes - [VERSIONS.md](https://xra1.com/VERSIONS.md): compatibility matrix across v1.0 / v1.1 / v2.0 - [BUILDING.md](https://xra1.com/BUILDING.md): how to use the reference Three.js adapter ## Authoring If you are asked to emit an XRAI scene: - Minimum viable: `{ "xrai_version": "1.0", "id": "", "scene": { "entities": [...], "relations": [], "events": [] } }` - Entity types: `object.primitive` (cube/sphere/cylinder/capsule/plane), `object.glb`, `object.hologram`, `object.light`, `object.emitter`, `object.wire-source`, `object.paint-stroke`, `object.parametric-stroke`; graph parts the viewer draws as labelled nodes laid out by their relations: `object.repo`, `object.dir`, `object.file`, `object.user`, `object.lang`, `object.page`, `concept` - Relation types (v1.0): `parent-of`, `wire-binds`, `reacts-to-audio`, `tracks` - Coordinate system: right-handed, Y-up, meters - Rotation: quaternion `[x, y, z, w]` - See [examples/](https://xra1.com/examples/) for 19 working fixtures ## Hand a person a world (no install) 1. Write a v1.0 document (above) and host it at any `https://` URL that serves it with CORS. 2. Give the person `https://xra1.com/viewer.html?scene=`. To open it on a chosen view, append `#view=,` (metres, the world's own coordinates) and optionally `&sel=` to outline one entity; the viewer's share button writes both. 3. They open it in any browser and can step inside it, see it in their room ("view in your room": AR Quick Look on iPhone, WebXR on Android and Quest; new, not yet tried on every device), share it or remix it. 4. Or give the person `https://xra1.com/viewer.html?from=`: it opens as a world of its parts and their relations (typing either URL into the viewer's box does the same). The world itself, as XRAI JSON for an agent: `https://xra1-world.xra1.workers.dev/from?url=` (built once, cached an hour, open to any origin). ## Query / MCP - Spatial MCP server: a proposal only, see [mcp-server/](https://xra1.com/mcp-server/); no server code is published here yet ## Runtimes - [runtimes/threejs](https://xra1.com/runtimes/threejs/): reference MIT adapter, the one engine the xra1.com viewer runs - [runtimes/](https://xra1.com/runtimes/): Unity (canonical engineering runtime), PlayCanvas, WebXR/Needle, Unreal, visionOS RealityKit — adapter stubs - [RUNTIMES.md](https://xra1.com/RUNTIMES.md): positioning (parallel MIT adapters, not blend; Icosa = Gallery/API, not a runtime) - [RUNTIMES_EVALUATION.md](https://xra1.com/RUNTIMES_EVALUATION.md): 9 objectives, 8 lock gates, honest scorecard — no runtime is LOCKED until evidence ## Governance - BDFL year 1 (@jamestunick), transfer to Apache SF or W3C Community Group once 1000+ external adopters - [CONTRIBUTING.md](https://xra1.com/CONTRIBUTING.md): contribution process - [CODE_OF_CONDUCT.md](https://xra1.com/CODE_OF_CONDUCT.md): short + direct - [rfcs/](https://xra1.com/rfcs/): RFC process + template + in-flight proposals (0001 events, 0002 hyperedges, 0003 conformance tests) - [SECURITY.md](https://xra1.com/SECURITY.md): responsible disclosure ## License - MIT for spec text + reference implementations + prompt libraries - CC0 (public domain) for the normative JSON schema - No CLA. No gatekeeper. No personal data collected; pages send anonymous view counts. ## Proof it ships Portals (reference implementation) — CVPR 2026 workshop paper on 4D World Models: - 60 FPS on iPhone 14 Pro (2.7–4.1× speedup vs prior art) - 360+ source-agnostic VFX on shared compute substrate - Stateful worlds (persistent across sessions, devices, users) - On-device voice intent parsing + cloud fallback + no-code authoring - iOS + web (Apple Vision Pro compatible) ## Repo - GitHub: https://github.com/imclab/xra1 - Discussions: https://github.com/imclab/xra1/discussions - Issues: https://github.com/imclab/xra1/issues - Web: https://xra1.com — @jamestunick (IMC Lab + H3M Inc.), 2026-04-22; worlds, viewer and runtime lines updated 2026-10-17