@geometra/renderer-webgpu

WebGPU renderer scaffold for Geometra.

This package provides:

• Capability detection via WebGPURenderer.isSupported()
• Async init() for adapter/device bootstrap (stub in first milestone)
• Stable renderer interface integration (render()/destroy())

Current status:

• API scaffold is implemented
• Full paint parity with @geometra/renderer-canvas is not implemented yet

Install

npm install @geometra/renderer-webgpu

Current support

• WebGPU capability detection and async device/context initialization
• Geometry render integration for solid box backgrounds
• 2D clear-pass fallback before init() completes
• Optional onFallbackNeeded(count) callback when a tree includes unsupported paint features

Current gaps:

• text paint
• selection highlight
• focus/debug overlays
• gradients, shadows, and border-radius parity
• full canvas feature parity

Usage

import { createApp, box, text } from '@geometra/core'
import { WebGPURenderer } from '@geometra/renderer-webgpu'

const renderer = new WebGPURenderer({ canvas })
if (WebGPURenderer.isSupported()) {
  await renderer.init()
}

await createApp(
  () => box({ padding: 16 }, [
    text({ text: 'WebGPU scaffold', font: '16px Inter', lineHeight: 22, color: '#fff' }),
  ]),
  renderer,
  { width: 400, height: 200 },
)

Notes

• Call await renderer.init() before relying on GPU rendering. Until then, render() uses a predictable clear-pass fallback.
• The current implementation is best treated as an MVP backend for capability probes and early integration work, not a canvas replacement.
• For the current backend matrix, see RENDERER_SUPPORT_MATRIX.md.