Cloudflare is developing a system that continuously evaluates whether verified bots follow website owners’ stated preferences. Instead of treating a verified identity as permanent proof of good behavior, the service observes how automated traffic acts over time and assigns it a trust band.
The initial signals include whether a bot respects robots.txt instructions and whether it reduces its request rate after receiving an HTTP 429 response. Consistently honoring those signals raises trust, while repeatedly ignoring them lowers it. Cloudflare says it looks for sustained patterns rather than punishing a bot for one isolated event.
The company is already using the system internally and plans to publish scores through Cloudflare Radar. It also intends to let customers choose the trust threshold used when handling verified-bot traffic. That would connect behavioral scoring with existing Bot Management controls, while Bot Preference Sync helps site owners publish their rules.
The approach could give publishers more information when deciding which crawlers or AI bots to permit. It is not yet generally available, and Cloudflare has not detailed a launch date or every signal that will affect a score. The system also depends on Cloudflare’s observations, so a trust band should be read as operational evidence rather than a universal certification.