Overview

luma.gl combines published, composable npm modules with private experimental packages under active development. Applications can choose the rendering, compute, and visualization functionality they need while checking the maturity of each package.

Use the API reference when you want the class-by-class reference pages for a specific module. If you are looking for conceptual guides or an introduction to how the pieces fit together, start in the API Guide and then come back here for the detailed type and method docs.

If you are looking for Model, start with @luma.gl/engine. The Model class lives in the engine module and is one of the main entry points for rendering in luma.gl.

Module	Usage	Description
@luma.gl/core	Required	The "Abstract" Device API (implemented by both the webgpu and webgl modules).
@luma.gl/webgl	Required *	Device adapter implemented using the WebGL API. Enables creation of WebGL resources.
@luma.gl/webgpu	Required *	Device adapter implemented using the WebGPU API. Enables creation of WebGPU resources.
@luma.gl/engine	Recommended	A set of WebGPU/WebGL independent core 3D engine style classes built on top of @luma.gl/core.
@luma.gl/scene	Experimental / Private	ANARI-inspired retained scene objects, instanced rendering, lights, materials, and HDR.
@luma.gl/shadertools	Recommended	Reusable shader modules, portable shader assembly, and application-defined shader hooks.
@luma.gl/effects	Optional	Composable post-processing effects, screen-space lighting, and reusable shader-pass pipelines.
@luma.gl/gpgpu	Optional	Portable GPU evaluation plus experimental gpu-data, gpu-core, and gpu-graph subpaths.
@luma.gl/arrow	Experimental / Private	Apache Arrow adapters for GPU layouts and GPU table objects from Arrow data.
@luma.gl/text	Experimental / Private	TextRenderer facade and caller-owned GPU text data.
@luma.gl/splats	Experimental / Private	Gaussian splat rendering and caller-owned prepared GPU splat data.
@luma.gl/experimental	Experimental / Private	Incubating APIs, including gpu-tables, rendering models, WebXR, and GPU analytics.
@luma.gl/gltf	Optional	Standards-first glTF assets, physical materials, character animation, and lossless interchange.
@luma.gl/test-utils	Optional	Test setups, in particular support for rendering and comparing images.

* At least one backend, either WebGL or WebGPU, must be installed to enable GPU resource creation.

Start Here

• @luma.gl/engine for Model, AnimationLoop, scenegraph helpers, and compute-oriented utilities.
• @luma.gl/scene for experimental, private ANARI-inspired declarative scenes, retained rendering, and instancing.
• @luma.gl/core for Device, buffers, textures, shaders, render passes, and RenderPipeline.
• @luma.gl/shadertools for shader modules and shader assembly.
• @luma.gl/effects for reusable image processing, bloom, and supported screen-space effects.
• @luma.gl/gpgpu for portable GPU evaluation, plus its experimental gpu-data, gpu-core, and gpu-graph subpaths.
• GPGPU data for primitive GPUData/GPUVector APIs, and Experimental GPU Tables for private GPURecordBatch/GPUTable APIs.
• @luma.gl/text for TextRenderer and GPU text data; use @luma.gl/arrow for Arrow conversion.
• @luma.gl/splats for experimental Gaussian splat rendering and caller-owned prepared GPU data.
• @luma.gl/experimental for incubating v10 APIs, including experimental WebXR frame, view, and raw camera helpers.
• @luma.gl/gltf for standards-first glTF assets, physical materials, character animation, native extensions, and source-faithful .gltf/.glb interchange.
• @luma.gl/webgl and @luma.gl/webgpu for backend adapters used by @luma.gl/core.
• @luma.gl/webgl/constants when you need raw numeric WebGL enums.