SceneView MCP

Give any AI assistant expert-level knowledge of 3D and AR development.

The officialModel Context Protocolserver forSceneView— the cross-platform 3D & AR SDK for Android (Jetpack Compose + Filament), iOS / macOS / visionOS (SwiftUI + RealityKit), and Web (Filament.js + WebXR).

Disclaimer:Generated code is provided "as is" without warranty. Always review before production use. SeeTERMS.mdandPRIVACY.md.

{ "mcpServers": { "sceneview": { "command": "npx", "args": ["-y", "sceneview-mcp"] } } }

Recommended — install theSceneView Claude Code pluginto get this MCP serverplus11 namespaced contributor commands and cross-platform reminder hooks in one shot:

/plugin marketplace add sceneview/claude-marketplace /plugin install sceneview@sceneview

claude mcp add sceneview -- npx -y sceneview-mcp

{ "mcpServers": { "sceneview": { "command": "npx", "args": ["-y", "sceneview-mcp"] } } }

Same JSON config as above. The server communicates viastdiousing the standard MCP protocol.

Every developer tool isfree: setup guides for every platform, code samples, the API reference, the migration tooling, the validator, model search, and the project analyzer.

get_best_practices·get_animation_guide·get_gesture_guide·get_performance_tips·get_material_guide·get_collision_guide·get_model_optimization_guide·get_web_rendering_guide·get_troubleshooting·debug_issue

search_models— find real 3D assets from the AI

Generated SceneView code is only useful if it points at an asset that actually exists.search_modelsqueries Sketchfab's public search API and returns a shortlist with names, authors, licenses, thumbnails, triangle counts, and viewer/embed URLs that the assistant can drop straight intorememberModelInstance(modelLoader, ...)or embed as a live preview.

Bring your own key (BYOK).SceneView never proxies the request — you keep the rate limit and the cost stays at zero. To set it up:
- Create a free account atsketchfab.com/register
- Copy your API token fromsketchfab.com/settings/password
- SetSKETCHFAB_API_KEYin your MCP client config:

{ "mcpServers": { "sceneview": { "command": "npx", "args": ["-y", "sceneview-mcp"], "env": { "SKETCHFAB_API_KEY": "YOUR_TOKEN_HERE" } } } }

Call it likesearch_models({ query: "red sports car", category: "cars-vehicles", maxResults: 6 }). If the key is missing, the tool returns a clear message explaining how to get one instead of failing silently.

generate_3d_model— create brand-new 3D assets from the AI

When no existing model fits,generate_3d_modelcloses the other half of the asset loop: it generates a fresh GLB from atext prompt(text→3D) or asource image(image→3D) via theTripo AIAPI, then returns a direct GLB download URL plus license/attribution metadata — ready forrememberModelInstance(modelLoader, ...)and AR placement.

Bring your own key (BYOK).Exactly likesearch_models: SceneView never proxies the request or holds your key — generations are billed toyourTripo account. To set it up:
- Create an API key atplatform.tripo3d.ai/api-keys(new accounts get free trial credits)
- SetTRIPO_API_KEYin your MCP client config:

{ "mcpServers": { "sceneview": { "command": "npx", "args": ["-y", "sceneview-mcp"], "env": { "TRIPO_API_KEY": "YOUR_KEY_HERE" } } } }

Call it likegenerate_3d_model({ prompt: "a low-poly cactus in a striped pot" })orgenerate_3d_model({ imageUrl: "https://example.com/chair.jpg", quality: "hd" }). Provide exactly one ofprompt/imageUrl.

⚠️ The GLB download URL expires ~5 minutes after generation— download the file immediately and self-host it (e.g. copy it into your app'sassets/models/). The tool result repeats this warning. Missing key, task failures, rate limits, and poll timeouts (2 min fast / 4 min hd cap) all return clear, actionable messages instead of hanging or crashing.

Because the MCP server runs on the user's machine,analyze_projectcan read their project files directly. Given apath(default:process.cwd()), it:

- Detects the project type by looking forbuild.gradle(.kts)withio.github.sceneview:sceneview(Android),Package.swiftwithSceneViewSwift(iOS), orpackage.jsonwithsceneview-web(Web).
- Extracts the SceneView dependency version and compares it against the latest release known to this MCP build, flagging outdated projects.
- Walks up to30source files (.kt,.kts,.swift,.js,.ts) and up to500 KBtotal, scanning for well-known anti-patterns: Filament/ModelLoader calls inside background coroutines, theLightNode(...) { ... }trailing-lambda bug, deprecated 2.x APIs (ArSceneView,TransformableNode,PlacementNode,ViewRenderable,loadModelAsync), andcom.google.ar.sceneform.*imports.
- Returns a structured{ projectType, sceneViewVersion, latestVersion, isOutdated, warnings, suggestions }report, plus a Markdown summary.

The tool is read-only, never writes to disk, and gracefully handles missing directories. Use it when the user asks "is my project up to date?" or as a quick sanity check before generating new code for an existing codebase.

The assistant callsget_ar_setup+get_sample("ar-model-viewer")and returns a complete, compilable Kotlin composable with all imports, Gradle dependencies, and manifest entries. Ready to paste into Android Studio.

"Create a 3D model viewer for iOS"

The assistant callsget_ios_setup("3d")+get_sample("ios-model-viewer")and returns Swift code with the SPM dependency, Info.plist entries, and a working SwiftUI view.

"What parameters does LightNode accept?"

The assistant callsget_node_reference("LightNode")and returns the exact function signature, parameter types, defaults, and a usage example — including the critical detail thatapplyis a named parameter, not a trailing lambda.

"Validate this code before I use it"

The assistant callsvalidate_codewith the generated snippet and checks it against 30+ rules: symbol existence against the real public API (unknown imports, made-up node types, nonexistent loader methods — each with did-you-mean suggestions), threading violations, null safety, API correctness, lifecycle issues, deprecated APIs. Problems are flagged with explanations before the code reaches the user.

Withoutthis MCP server, AI assistants regularly:

- Recommend deprecatedSceneform(abandoned 2021) instead of SceneView
- Generate imperativeView-basedcode instead of Jetpack Compose
- Usewrong API signaturesor outdated parameter names
- Miss theLightNodenamed-parameter gotcha (apply =not trailing lambda)
- Forget null-checks onrememberModelInstance(it returnsnullwhile loading)
- Have no knowledge of SceneView's iOS/Swift API at all

- Always use the current SceneView 4.0.x API surface
- Generate correctCompose-native3D/AR code for Android
- Generate correctSwiftUI-nativecode for iOS/macOS/visionOS
- Know about all 46+ node types and their exact parameters
- Validate code against 30+ rules before presenting it
- Provide working, tested sample code for 33 scenarios

The MCP server is tested with1,965 unit testsacross 83 test suites covering:

- Every tool response (correct output, error handling, edge cases)
- All 33 code samples (compilable structure, correct imports, no deprecated APIs)
- Code validator rules (true positives and false-positive resistance)
- Node reference parsing (all node types extracted correctly fromllms.txt)
- Resource responses (API reference, GitHub issues integration)

Test Files 83 passed (83) Tests 1965 passed (1965)

All tools workfully offlineexceptsceneview://known-issues(GitHub API, cached 10 min),search_models(Sketchfab, BYOK), andgenerate_3d_model(Tripo AI, BYOK).

"MCP server not found" or connection errors

Install Node.js fromnodejs.org(LTS recommended). npm and npx are included.

{ "mcpServers": { "sceneview": { "command": "npx", "args": ["-y", "sceneview-mcp"], "env": { "HTTPS_PROXY": "http://proxy.example.com:8080" } } } }

A small set of domain-specific tools is gated behind an optional subscription. They aren't required for general 3D/AR work — only useful if you happen to be building one of these specific verticals:

- Automotive— car configurator, paint shader, parts catalog, HUD overlay, AR showroom
- Gaming— physics, particles, level editor, character viewer, inventory 3D
- Healthcare— surgical planning, dental viewer, medical imaging, anatomy, molecule viewer
- Interior— room planner, lighting design, material switcher, furniture placement, room tour

Plus 3 generation helpers:render_3d_preview,create_3d_artifact,generate_scene.

The base SDK and every developer tool listed above stay free, always.

If sceneview-mcp saves you time, considersponsoring on GitHub Sponsors. Building this is a one-dev labor of love and donations keep the free tier covered.

Enabled by default on the free tier (MCP client name/version and tool names — no personal data, no prompt content). Opt out withSCENEVIEW_TELEMETRY=0. SeePRIVACY.mdfor the full payload shape.

cd mcp npm install npm run prepare # Copy llms.txt + build TypeScript npm test # 1965 tests npm run lint # Biome (repo-root biome.json) — lint + format + import assists npm run lint:fix # same, applying the safe fixes npm run dev # Start with tsx (hot reload)

mcp/ src/ index.ts # MCP server entry point tools/handler.ts # Tool dispatcher (free + pro) tiers.ts # Free vs Pro tier mapping samples.ts # 33 compilable code samples (Kotlin + Swift) validator.ts # Code validator (30+ rules) node-reference.ts # Node type parser guides.ts # Best practices, AR setup, roadmap, troubleshooting migration.ts # v2 -> v3 -> v4 migration guide preview.ts # 3D preview URL generator artifact.ts # HTML artifact generator (model-viewer, charts, product 360) issues.ts # GitHub issues fetcher (cached) search-models.ts # Sketchfab BYOK search generate-model.ts # Tripo BYOK text/image -> GLB generation analyze-project.ts # Local project scanner proxy.ts # Pro-tool proxy to hosted gateway llms.txt # Bundled API reference (copied from repo root)

- LICENSE— MIT License
- TERMS.md— Terms of Service
- PRIVACY.md— Privacy Policy

An MCP server tailored for React Native–first development using Gluestack UI