Donor-database site blocks Amazon's AI crawler after logging 117,000 daily requests
PatronView, a database of U.S. museum and cultural-institution donor records, blocked Amazon’s Amzn-SearchBot crawler in early August after founder Nick Gray logged roughly 117,000 requests from it in a single day, he said in published server-log findings. The site had already blocked Anthropic’s Claude-SearchBot in June after measuring a 35,000-to-1 ratio of crawls to human visitors referred; a 48-hour CAPTCHA challenge in the same period drew 106,437 attempts and only 252 solves, a 0.24% rate, per the log data cited by PPC Land.
Why it matters: It's a rare, publicly documented data point on how disproportionate AI-crawler traffic can run relative to referral value, and it shows the kind of math behind operators moving from monitoring to outright blocks ahead of Cloudflare's September 15 default block.
The record: Anthropic
Glossary: AI crawlers
Via PPC Land ↗
Posted to the wire August 8, 2026.