Siteline
Benchmark
Siteline keeps a small, curated benchmark set for long-term comparison: major public sites plus agent-relevant infrastructure projects. This page publishes the target set and recent public sample results.
The current target list contains 50 domains. It is intentionally small, stable, and scan-suitable so week-to-week changes are easier to interpret.
Machine-readable target list:
/data/benchmark-domains.json
The set is not an endorsement list and it is not an industry average. It is a public calibration corpus for scanner behavior, examples, and long-term trend analysis.
Example grades from the benchmark corpus (April 2026 pass): google.com C 68, publedge.org C 77, everyailaw.com A 97, prompterkit.app D 63, github.com F 44. Stored reports expire after 30 days, so these examples are shown without links; run a fresh scan from the homepage to see a current report for any of these domains.
The public benchmark page is the first step. Real industry baselines require a dedicated history table, source tagging, enough examples per industry, and a transparent method for excluding blocked or unsuitable targets.