Small-footprint traffic measurement.

We measure whether the public site is reaching people, search engines, and AI systems. The tracker is first-party and designed to retain the minimum useful data.

What is retained

Request time, page path, coarse country code supplied by the hosting edge, referring hostname only, UTM campaign labels, an anonymous session hash, broad device/browser family, and a traffic class such as likely human, search crawler, AI crawler, AI assistant, link preview, monitor, automation, or unknown.

What is not retained

No raw IP address, no full user-agent string, no full referring URL, no form content, no wallet address, no cross-site advertising identifier, and no browser fingerprint.

How classification works

Known crawler signatures are checked first. A browser that completes the small first-party client signal can be marked “confirmed human”; otherwise ordinary browser-shaped requests remain “likely human.” These are useful estimates, not claims of perfect identity.

Retention

First-party request and interaction records are automatically limited to a rolling 90-day window. Vercel’s separate privacy-oriented Web Analytics may also process basic aggregate page-view data under its platform terms.

Contact

Questions can be sent to security@sovxus.com.