site analytics · live
These are PIGEON's real numbers — the same data the private operator dashboard computes, published as aggregates only. There are no IP addresses, no per-client rows and no user-agent strings anywhere on this page or in the API behind it: the public endpoint copies fields through an explicit allowlist, so what you see is every field that is allowed out, not a sample of a larger feed.
Loading from GET /v1/stats…
the honest headline
how the traffic splits
Traffic is separated by source. "Self" is the host's own public IP or user-agent; "operator-test" is the operator's own first-contact test VPS range. Neither is a customer, and neither is ever counted as demand.
the scanner skew
sessions
A session is a client's requests grouped with a 30-minute gap. Medians matter more than means here, because one scanner drags every average up.
discovery funnel
| stage | clients | requests | % of entering |
|---|---|---|---|
| Loading… | |||
settlement
trends
| day | req | clients | 402s | settled |
|---|
| hour | requests |
|---|
| hour (UTC) | requests | clients | 402s | settled |
|---|
| day | new | returning | active | requests |
|---|
where they came from
| referrer host | count | class |
|---|
page engagement
| path | views | clients | entry | exit |
|---|
discovery artifacts
| path | hits | clients | status |
|---|
Machine-facing files are where an agent-based visitor is really found. They are read far more than the human pages.
caveats
Raw JSON: /v1/stats (aggregates) · /v1/traffic (who knocked, separately) · OpenAPI · llms.txt. This page renders only fields the endpoint allowed out.