Sources
webrank is a composite of eight public sources plus the Common Crawl web graph — no single list decides anything. Each source is fetched on its own cadence, normalized to pay-level domains, and weighted by how well it agrees with real-traffic data (see the methodology).
| source | what it measures | coverage | cadence | license |
|---|---|---|---|---|
| Tranco | research-grade popularity composite | 4.5M | daily | free, citable |
| host.io | web + DNS popularity | 10.0M | monthly | free tier, attributed |
| Open PageRank | PageRank over Common Crawl links | 6.9M | monthly | free, attributed |
| Chrome UX Report (CrUX) | real Chrome user visits | 759k origins | monthly | open (Google) |
| Cloudflare Radar | DNS resolver traffic | 1M | daily | CC-BY |
| Majestic Million | referring subnets (links) | 1M | daily | free, attributed |
| Cisco Umbrella | DNS resolver traffic | 226k PLDs | daily | free |
| Common Crawl web graph | inter-domain links (~140M domains, billions of edges) | ~140M | ~monthly release | open |
Coverage counts are live pay-level-domain counts from our latest load (they shift slightly per release). Two things worth knowing:
- Different sources measure different things. DNS lists (Umbrella, Radar) see resolver
queries — CDNs and API backends rank high. Browser data (CrUX) sees what people actually visit. Link data (Common Crawl, Majestic, Open PageRank) sees authority, not visits. That's exactly why webrank publishes two numbers: popularity and authority.
- Coverage shapes the rank. Traffic sources reach a few million domains deep; only the
link graph covers the long tail. See coverage for what the rank means at each depth.
We redistribute per-source ranks as an attributed cross-reference in the API (ranks field) — every source above permits redistribution under the listed terms. Sources that don't allow it would still count in the composite but be omitted from ranks; today none are omitted.
Related: methodology · coverage · webrank vs Tranco