Live scanner metrics

Grade distribution

How many of the tracked hosts currently sit at each grade. The spec is heavy: even mid-tier devtool sites land in C or D if they do not publish an explicit AI-bot section in robots.txt plus an llms.txt.

Score distribution (0-100, bucketed)

Where the mass of hosts sits on the 0-100 scale. Each bar is a 10-point bucket. A spike at the low end usually reflects sites that score the minimum 6 points for a reachable robots.txt and nothing else. A spike at the high end reflects sites that have invested in llms.txt + structured data + AI meta directives together.

Top improvers (last 30 days)

Hosts that moved up the most since the first scan logged in the last 30-day window. Delta is from first scan to most recent scan. A positive delta means the site's public signals improved in a way our rubric rewards.

Site health

Live output from the 15-minute ping monitor that probes eight critical surfaces (homepage, /api/health, AI Readiness, live leaderboard, llms.txt validator, pricing, analysis hub, stats). Uptime is measured externally from the same machine — see the caveat note below.

Note: the pinger runs on the same host the site runs on, so this measures application availability (did /api/health return 200?), not network reachability from the outside world. For an independent uptime check, any third-party monitor will do.