Abuse network

Enable Abuse network in site settings, or PATCH /api/sites/:key with {"network":true}. It is off by default. Sites that opt out neither contribute nor receive matches.

Reports and matches

Fraud feedback, cluster Fraud labels and flagged usage contribute keyed hashes from verification events recorded while the site was opted in. A later verification gets network_reported with detail: owners=N only if two distinct other customer owners reported a matching key in the last 90 days. Your own reports never count toward your matches. Multiple sites belonging to one owner count once.

The match is an abuse reason rather than a bot score. Enforcement follows the site's mode and rules. Observe mode returns allow with its would-have decision. The overview shows Network matches for the selected 24h, 7d or 30d range.

Privacy and limits

The network stores only global keyed hashes of device fingerprints, normalized emails and non-freemail email domains. Weak fingerprints are combined with their IP-prefix hash, as in local actor matching. Raw fingerprints, raw emails, raw domains, IP addresses and site-local visitor ids are never shared. Verification event features contain networkKeys only for opted-in sites.

Each owner may contribute at most 1,000 report-key rows in a rolling day. Reports expire after 90 days and are pruned by scheduled cleanup. Demo, test and ownerless sites cannot participate. During storage outages, verification continues without network matches.

Include this optional fraud-prevention processing in your privacy notice. See the privacy policy and DPA.