3D & AR SDK for Android, iOS, Web — API docs, samples, validation, and code generation.
io.github.sceneview/mcp (MCP) Server
This MCP server is associated with SceneView, a “3D & AR SDK for Android, iOS, Web” that provides API documentation, samples, validation, and code generation. It is described in the repository as supporting multiple platforms and frameworks using shared concepts across targets.
🛠️ Key Features
3D & AR SDK for Android, iOS, and Web
API docs
Samples
Validation
Code generation
🚀 Use Cases
Building 3D and AR experiences on Android, iOS, and Web
Using existing UI frameworks across platforms (Android, iOS, Web, Desktop, TV)
Supporting Flutter and React Native workflows
⚡ Developer Benefits
Reuse the same concepts across platforms
Generate code from the server’s provided capabilities
Access documentation, samples, and validation materials
⚠️ Limitations
The provided information only specifies Android, iOS, and Web, plus additional platforms mentioned in the excerpt, without further operational details.
The AI-first 3D & AR SDK for Jetpack Compose — and the web.
Declarative 3D and AR for app developers who want a model on screen, or in the room, without a
game engine: Filament + ARCore on Android, Filament.js + WebXR in the browser, and a Compose
Multiplatform viewer. It is the
maintained successor to Google Sceneform. Open source, Apache 2.0.
Used in production. More than 1,300 stars on GitHub, and over 500 public repositories declare
SceneView as a dependency. It ships in Medal, IKEA, Lenskart and Safie on Google Play, is declared by the
open-source apps Bitkey (Block) and Vocable (WillowTree), and is a dependency of the GeneXus
Augmented Reality module and the TruVideo camera SDK. The Niantic Spatial SDK samples use it for AR.
AI-first means one thing here: an assistant that reads llms.txt or
the MCP server writes SceneView code that works on the first try. When it doesn't, the
API or the doc gets fixed.
The demo apps are built from samples/. Any demo opens straight from a link:
https://sceneview.github.io/open?demo=<id>. Start with
Cosmos, the galaxy the catalogue opens on, or
3D Scan, a space scanned with ARCore and rebuilt in 3D.
Install the Android demo, record your screen while you use a demo, and tell us what breaks or
confuses you: open an issue
with the recording attached, or send it from the app (any demo, Settings, then Send feedback).
Quick start
One minimal, working example per platform. The full reference for each is in
llms.txt.
The app module needs compileSdk = 37, minSdk 24 or higher, AGP 9.4.0+, Kotlin 2.3 or newer and JVM
target 21 for both Java and Kotlin: SceneView ships Java 21 bytecode with public inline
functions. The quickstart requirements give the Gradle
snippet and the reason for each.
Stuck on AGP 8 / compileSdk 36 / Kotlin 2.2? Use 4.18.0, the newest release that builds there.
models/helmet.glb lives in src/main/assets/. rememberModelInstance returns null until the
model is loaded, then recomposes. Never call modelLoader.createModel* from a background
coroutine: Filament calls must run on the main thread, and rememberModelInstance handles that.
AR (Android)
kotlin
// build.gradle.kts
implementation("io.github.sceneview:arsceneview:4.69.5") // includes the 3D module
Plane detected → anchor set → Compose recomposes → the model appears. AR state is just Kotlin
state. For placement without a tap, grounded and gesture-ready, start from
AutoPlacementScene (see llms.txt).
arsceneview already declares the CAMERA permission and marks ARCore as optional. An AR
Optional app declares android.hardware.camera.ar and android.hardware.camera as
<uses-feature … android:required="false" />; an AR Required app overrides the ARCore value
with tools:replace="android:value". Both snippets are in the
AR manifest section.
A viewer by design — load, orbit, light, tap; no AR. Scope and the one-time iOS setup:
sceneview-compose/.
iOS, Flutter and React Native
SceneView for SwiftUI (SceneViewSwift), Flutter (flutter_sceneview) and React Native
(@sceneview-sdk/react-native) is deprecated: 4.68.0 is their last release. Published versions stay
installable and the source stays in this repository — what that means and the timeline.
Shipped means released and publicly reachable; the second word is the API maturity.
The other platforms in the table below join this section once a capture backs them.
Everything an assistant needs to write SceneView code ships with the SDK:
llms.txt — the complete API reference in one file: composables, every node
type, threading rules, recipes. Its Kotlin snippets are compiled in CI. Served at
https://sceneview.github.io/llms.txt for tools without MCP support.
Rules files — AGENTS.md (Codex, Cursor, GitHub Copilot, Gemini in Android Studio and
others), CLAUDE.md (Claude Code), .github/copilot-instructions.md, and .cursorrules for
older Cursor versions.
The MCP server — free, no API key. The tools assistants reach for most:
validate_code (checks a snippet for known pitfalls and unknown symbols before you run it; it does not check argument names),
get_node_reference (the exact node signature, not an invented one), list_samples /
get_sample (38 samples to start from), and get_setup / get_ar_setup (project wiring).
// Cursor (.cursor/mcp.json), Cline, JetBrains AI Assistant{"mcpServers":{"sceneview":{"command":"npx","args":["-y","sceneview-mcp"]}}}// VS Code (.vscode/mcp.json) uses the "servers" key instead{"servers":{"sceneview":{"type":"stdio","command":"npx","args":["-y","sceneview-mcp"]}}}
Clients that only speak HTTP (Gemini in Android Studio, ChatGPT) use the hosted endpoint
https://mcp.sceneview.dev/mcp, or run their own with npx sceneview-mcp --http. Per-client
snippets: sceneview.github.io/#ai-setup. Listed on the
MCP Registry.
ChatGPT / Codex plugin
This repository is also an OpenAI plugin: .codex-plugin/plugin.json points at the three
skills under agents/ (sceneview, sceneview-ios, sceneview-web). From a
checkout:
bash
codex plugin marketplace add "$PWD"# absolute path — a relative one does not resolve
codex plugin add sceneview@sceneview-local
Codex also discovers the skills from .agents/skills/ on its own. Over HTTP the MCP server
carries an inline view_3d_model widget that renders a public GLB/glTF URL in the
conversation. Listing copy and test prompts: agents/OPENAI-PLUGIN.md.
Claude Code plugin
/plugin marketplace add sceneview/claude-marketplace, then /plugin install sceneview@sceneview,
installs the same three skills as the Codex plugin together with the MCP server. The commands used
to work on this repository are a separate plugin, sceneview-contrib@sceneview — see
sceneview/claude-marketplace.
Vertical MCP servers (Rerun AR debugging and others) are listed in the MCP README.
Android in depth
3D scene
SceneView is a composable that renders a Filament viewport. Nodes are composables inside it.
Change state → the tree updates. No imperative parent.addChild().
SceneScope / ARSceneScope DSL
Model formats
Format
Android
Apple
Web
glTF / GLB
✅
— convert to USDZ (tools/convert-usdz.sh)
✅
USDZ / Reality / USD
—
✅ RealityKit
—
STL (binary + ASCII)
✅ converted to GLB in memory
✅ ModelIO, millimetres by default
—
OBJ + MTL
✅ converted to GLB, material colours
✅ ModelIO
—
PLY (binary + ASCII)
✅ converted to GLB, vertex colours
✅ ModelIO, vertex colours
—
3MF
✅ converted to GLB, declared units honoured
✅ SceneViewSwift parser, declared units honoured
—
There is no per-format API on Android: rememberModelInstance(modelLoader, path) accepts all of
them. On Apple, RealityKit reads no glTF, so .gltf / .glb are deliberately not in
ModelFormat; USD/Reality load through RealityKit and the mesh formats through MeshAsset. The format is decided by the file's bytes, not its extension, so a file shared into your
app as application/octet-stream with no name still opens. The converters are dependency-free
Kotlin in sceneview-core. Tracked next:
one ModelFormat entry point and
every format on the web. Details:
Model formats.
Apple (iOS / macOS / visionOS)
Native Swift Package built on RealityKit, with a node set mirroring the Android API. Content you
can build synchronously uses the @NodeBuilder form:
Two <script> tags and one call (see Web above). The wrapper is ~24 KB gzipped; the
engine it drives is Filament — the same renderer as Android SceneView — compiled to WebAssembly
(~2.3 MB gzipped).
JavaScript API (script tag):
SceneView.modelViewer(canvasOrId, url, options?) — all-in-one viewer with orbit + auto-rotate
SceneView.create(canvasOrId, options?) — empty viewer, load a model later
viewer.loadModel(url) — load or replace a glTF/GLB model
viewer.setAutoRotate(enabled) — toggle rotation
viewer.dispose() — release resources
Kotlin/JS (sceneview-web, npm only) — the power-user API: OrbitCameraController, the
geometry DSL, reactive node updates, and WebXR through ARSceneView (immersive-ar,
hit-test, anchors, light estimation), VRSceneView (immersive-vr, controllers) and the
low-level WebXRSession. The module builds a webpack bundle, so it has no Maven coordinate:
bash
npm install sceneview-web
The package expects a Filament global and does not include the SceneView.modelViewer script
helpers. A Kotlin Multiplatform project that only needs the shared core (collision, math,
geometry, animation, physics — no renderer) uses
implementation("io.github.sceneview:sceneview-core-js:4.69.5").
Google archived Sceneform in 2021 and ships
no first-party declarative AR renderer. SceneView descends from the maintained Sceneform
community fork: ARCore for perception, Filament for rendering, Jetpack Compose for the API, and
glTF (.glb / .gltf) instead of the deprecated .sfb format.
Coming from Sceneform? The migration guide maps it
concept by concept (ArFragment → ARSceneView { }, ModelRenderable →
rememberModelInstance, and so on).
AR debugging
Record & Replay — capture an ARCore session once with rememberARRecorder(), replay it 1:1
at your desk with ARSceneView(playbackDataset = file). See
docs/docs/ar-recording.md and the
Record & Playback demo.
Hosted Rerun viewer — export a scan's Rerun recording from the 3D Scan demo, host the .rrd file at
an https address that ends in .rrd, and open
https://sceneview.github.io/rerun/?url=<encoded-url> to scrub the session frame by frame in a
browser, with no local install. The host has to allow cross-origin requests (CORS): a public
GitHub repository's raw.githubusercontent.com link and GitHub Pages do, a GitHub release
asset does not (checked 2026-10-13); an R2 or S3 bucket does with a CORS rule that allows
https://sceneview.github.io. Architecture and the Kotlin API
(RerunBridge.requestSaveAndShare) are in the AR Debug — Rerun.io section of
llms.txt.
Architecture
Each platform uses its native renderer. Shared logic lives in Kotlin Multiplatform.
SceneView runs in commercial products. Sponsor it.
Commercial products ship SceneView. Almost nobody contributes back, in code or in funding, while the maintainers carry every fix and every release. If SceneView is part of your product, sponsor it.
Sponsor. Funds the fixes, Filament upgrades and releases your product depends on. Open Source Collective issues the invoice and the receipt; your company is listed nowhere unless you ask.