27,067
domains parsed this edition
27,067
per-domain rows published
| Item | Value |
| Edition generated | 2026-08-09T10:47:13.783Z |
| Age | 1.5 days (within the weekly cycle) |
| Methodology version | 2026-08-10.1 |
| History span | 2026-08-08 → 2026-08-09 (2 points) |
| Crawlers tracked | 20 |
| Publisher panel | 49 domains |
| Cadence | Weekly full sweep, automated |
Enforcement sample — read this before quoting it
Current edition: 50% of declared blocks were actually enforced, across 24 checks on the publisher panel.
This figure moves between runs, and you should know why. It is measured on a small panel (tens of checks, not thousands), and the denominator changes with how many panel domains responded to an identified crawler in that sweep. We have observed it between 50% and 62% on consecutive editions. Treat it as directional evidence that a substantial minority of declared AI blocks are not enforced — consistent with independent findings of roughly 40% — and always quote it with its n and date. It is not a precise population estimate, and we will not present it as one.
What we know is imperfect
- Declaration is not enforcement. Headline block rates measure what sites publish, not what they do.
- One vantage point. We crawl from a single network; sites that vary by geography or ASN may look different elsewhere.
- Prices are sparse. Very few sites quote a machine-readable price. We report the count rather than smoothing it.
- Not every domain answers. Rates are always computed against domains parsed, never the nominal frame.
- ccTLD is a proxy for country, not a measure of publisher nationality or audience.
Corrections
No corrections have been issued. When one is, it appears here and in the machine-readable corrections array of the dataset, with the date, what changed and why. We do not silently amend past editions.
Verify any of this yourself
curl -s https://api.crawlpriceindex.com/v1/check?domain=context
curl -s https://api.crawlpriceindex.com/v1/methodology
curl -s https://crawlpriceindex.com/index.json