FAQ
Is it really free? What's the catch?
Yes — no API key, no signup, data under CC-BY 4.0 (attribute "webrank.top"). We build domain intelligence commercially; the open rank is the foundation layer and publishing it openly makes it better (more eyes, more citations, more trust). Rankings are never for sale — see about.
Why should I trust webrank over Tranco or a commercial score?
Don't trust — check. The method is published including the weight derivation and validation numbers (+46% over equal weighting at predicting held-out real traffic), every API response includes the per-source ranks so you can verify the composite, and monthly snapshots are permanent so any claim can be re-tested. Full comparison: webrank vs Tranco.
Why does my domain have a low rank / no rank?
The rank covers domains seen by at least one of the eight sources. A new or small site typically enters via the Common Crawl link graph first (someone links to it), then climbs as traffic sources pick it up. Deep ranks are link-driven — see coverage for what a rank means at each depth.
webrank vs pagerank — which number do I use?
webrank answers "how much is this domain used?" (traffic, popularity rank).
pagerank answers "how much does the web vouch for it?" (links, 0–10). Fraud checks: use both — mismatches are the signal. Explainers: popularity, authority.
How fresh is the data, and can I cite it?
The rank recomputes daily (~04:00 UTC) from the daily sources; monthly sources land on their own cadence. On the 15th we freeze a permanent snapshot (webrank-YYYY-MM.csv.gz) — cite those; they never change. Details: updates, cite.
Can a domain buy a better rank? Can spammers game it?
No one can buy a rank (including advertisers, if we ever have any). Gaming is expensive by design: the composite spans eight independent sources, link spam moves pagerank inputs but not traffic, bot traffic moves DNS sources but not links or Chrome data — and the two-number design makes the manipulation itself visible. See manipulation resistance.
Are there rate limits?
Be reasonable (cached GETs are effectively unlimited; batch with POST /rank instead of hammering). For bulk work, download the list — it's the same data.
How do I get all 146M domains, not just the top-1M?
The full list ships as split archives with each monthly snapshot — see the webrank-lists GitHub repo. The API serves any domain individually regardless.