# webgpu.report

> Aggregate WebGPU capability statistics collected from real machines: feature availability, limits, hardware/browser mix, and long-term trends. Developer-leaning sample; metric formulas documented at /about.

Machine-readable data lives under /api/v1 (JSON, cached, rate-limited, CORS-open; CSV via format=csv on time-series endpoints). OpenAPI spec: /api/openapi.json.

## Key endpoints
- /api/v1/stats/summary: totals, WebGPU availability %, OS/browser/GPU-vendor shares
- /api/v1/stats/features?feature=subgroups&bucket=month: availability time series per feature
- /api/v1/stats/breakdown?dimension=vendor: share-by-dimension time series
- /api/v1/stats/limits?name=maxBufferSize: limit percentiles (p50/p90/p99)
- /api/v1/features: feature dictionary

## Pages
- /features: feature availability index
- /features/<name>: per-feature explanation and trends
- /about: methodology and caveats

Please cache responses and honor Retry-After on 429s. Heavy sustained use: request an API key (see /api).