An observatory, not a fortress

Who runs it

Botservatory is built by William Mushkin, who also runs who.is, time.io and dns.pizza.

How the live data works

Each request to this site is checked against the catalogue. If a request claims to be a listed crawler, its address is tested against the operator's published file, then reverse DNS where the operator supports it. The result is one of four things: verified by list, verified by DNS, failed, or impossible to verify because the operator publishes nothing to check against.

A failed check means the request did not match what the operator publishes. That usually means an impersonator, but a newly added range can also fail until the file catches up.

Web Bot Auth signatures are noted when presented, but not yet validated, and the logbook labels them that way. Requests with no bot markers are only counted. Stored events contain no IP address, whole or hashed.

What is not live

The guides, the catalogue's descriptions, the numbers and the ephemeris are editorial and show their dates. Nothing on this site is simulated. If an instrument has no data, it says so.

This site's own bot policy

We want to be read, so our robots.txt allows search, fetchers and training on our editorial pages, and asks training crawlers to stay out of the tools and API. We send Content-Usage: train-ai=y to match, so what we ask for and what we allow agree. Nothing is blocked at the firewall, because this site exists to observe.

The catalogue data is available under CC BY 4.0. Please cite the original sources for the numbers, because they did the measuring.