WebGPUbridge

A small, explicit bridge from portable 3D descriptions to WebGPU.

shaderlab.dev

WebGPUbridge materializes Web3D Kit descriptors as WebGPU-ready data and caller-owned GPU resources. It keeps scene descriptions portable while leaving adapter selection, device ownership, rendering, and asset loading explicit.

Install

npm install webgpubridge web3d-kit

import { mesh, planeGeometry, shaderMaterial } from "web3d-kit";
import { createWebGPUObject } from "webgpubridge";

const object = createWebGPUObject(
  mesh({
    geometry: planeGeometry({ width: 2, height: 2 }),
    material: shaderMaterial({ vertexShader, fragmentShader }),
  }),
);

Architecture

shader-runtime        webgl-kit
      │                  │
      └─────────┬────────┘
                ▼
            web3d-kit
                │
                ▼
           webgpubridge
                │
                ▼
              WebGPU

web3d-kit owns portable descriptions; webgpubridge materializes those descriptions without taking ownership of a WebGPU adapter, device, or queue.

Supported conversions

• vectors, transforms, bounds, scenes, groups, meshes, and cameras
• plane, box, sphere, and simple position/index buffer geometry
• basic, standard, and shader material envelopes
• tightly packed scalar and vector uniform blocks
• explicit GPU buffer, shader module, and render-pipeline helpers

Lifecycle

Use disposeWebGPUBuffer or disposeWebGPUGeometry explicitly. The bridge does not request an adapter or device, submit work, or dispose caller-owned pipelines and textures.