AI WebGPU Lab Template

Minimal WebGPU Starter

`tpl-webgpu-vanilla` is the thinnest raw WebGPU entry point in the org. It acquires an adapter and device, configures a canvas surface, renders an animated triangle, and exports a schema-aligned JSON draft you can promote into reports/raw/.

This page is intentionally small enough to copy into downstream experiments before framework or runtime-specific code lands.

Starter Controls

Run the capability probe first. If WebGPU is available, start the triangle sample to capture init and frame pacing metrics.

Template Guidance

  • Use this surface when a new repo needs the smallest possible WebGPU success path.
  • Keep capability and fallback reporting intact before layering renderer abstractions.
  • Promote the exported JSON into reports/raw/ only after validating browser, OS, and adapter details.

Canvas Sample

Metrics

Environment Snapshot

Activity Log

    Schema-Aligned Result Draft

    {
      "status": "pending"
    }