webgpujs

library0.0.30jsnpmunverified

WebGPUjs (v0.0.30) is a JavaScript library that enables writing full-featured WGSL pipelines directly in JavaScript, without requiring separate shader files. It provides a high-level API for allocating buffers, uniforms, textures, and specifying compute, vertex, and fragment shaders. Key features include automatic binding combination, support for compute shaders with multiple output buffers, chaining of shader stages, and instance VBOs. As an early-stage project (pre-v1), it breaks often and lacks comprehensive docs. It competes with raw WebGPU API and wgpu-native bindings, focusing on developer ergonomics for rapid prototyping. Release cadence is irregular; expect breaking changes.

testing

npm install webgpujs

package ↗

INSTALL

IMPORT

SIG · WEBGPUJS

webgpujs

testingjavascriptv0.0.30

harness data pending

Install & Compatibility

Where this runs

No compatibility data collected yet for this library.

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

GPU

✓ import { GPU } from 'webgpujs'

✗ const GPU = require('webgpujs')

ESM-only; package does not support CommonJS require. TypeScript types included.

ComputePipeline

✓ import { ComputePipeline } from 'webgpujs'

✗ import { computePipeline } from 'webgpujs'

Exported class names are PascalCase. Lowercase import will fail.

TranspileWGSL

✓ import { TranspileWGSL } from 'webgpujs'

Utility for converting JS-like WGSL syntax; not always needed.

type ComputePipelineOptions

✓ import type { ComputePipelineOptions } from 'webgpujs'

✗ import { ComputePipelineOptions } from 'webgpujs'

Interface only; use type import to avoid runtime import errors.

Creates a WebGPU compute pipeline that doubles each element of a 1024-element float32 array using a WGSL shader.

import { GPU, ComputePipeline, Buffer } from 'webgpujs'; async function run() { const gpu = await GPU.create(); const pipeline = new ComputePipeline(gpu, { code: ` @group(0) @binding(0) var<storage, read_write> data: array<f32>; @compute @workgroup_size(64) fn main(@builtin(global_invocation_id) id: vec3<u32>) { data[id.x] = data[id.x] * 2.0; } `, buffers: [{ name: 'data', size: 1024 * 4 }], workgroupCount: [16, 1, 1], }); const data = new Float32Array(1024); data.forEach((_, i) => data[i] = i); const buffer = await pipeline.run({ data: data }); await gpu.dispatch(pipeline); const result = await buffer.read(); console.log('Result:', result); } run().catch(console.error);

Debug

Known issues

breakingGPU.create() was previously synchronous; now returns a Promise.

fix

Use await GPU.create() or .then().

affects: >=0.0.20

deprecatedBuffer type strings (e.g., 'storage') replaced with buffer descriptor objects.

fix

Replace buffer type strings with { type: 'storage', ... }.

affects: >=0.0.25

breakingpipeline.run() no longer returns a buffer array; returns a single buffer object.

fix

Access result via .read() on the returned buffer.

affects: >=0.0.28

gotchaWGSL shader code must be inside compute shader functions; bare code not supported.

fix

Wrap code in an entry point function (e.g., @compute @workgroup_size(64) fn main(...)).

affects: all

gotchaBuffer size must be aligned to 4 bytes (f32). Unaligned sizes cause errors.

fix

Ensure buffer sizes are multiples of 4.

affects: all

Errors

Common errors & fixes

TypeError: GPU.create is not a constructor

Using 'new GPU.create()' instead of 'GPU.create()' (static method).

fix

Replace with 'const gpu = await GPU.create();'

TypeError: Cannot read properties of undefined (reading 'read')

pipeline.run() returns undefined if pipeline is not properly configured or if WebGPU device is lost.

fix

Check that pipeline code is valid and that GPU device is still available.

WGSL syntax error: expected 'fn' keyword

Missing function declaration in shader string.

fix

Wrap your compute logic in a function with @compute attribute.

Error: Requested device adapter not found

No WebGPU-compatible GPU or browser not supporting WebGPU.

fix

Run in a recent Chrome/Edge or Firefox nightly; ensure WebGPU enabled.

Upgrade

Version history

0.0.30latest on npm

Audit

Dependencies

No dependency data recorded yet.

Agent activity

14 hits · last 30 days

node

Amazon

Resources

packagewebgpujs ↗