The Directory now answers questions, and hands out its data
The Directory went up this morning as one place to browse what Skydex knows. The second half of that work went out this evening, and it is mostly for people who arrive from a search box or an assistant rather than from our menu.
Questions on every page
Each Directory section and each subject page now ends with a short Questions block. The answers are built from the page's own numbers: which aircraft type the network hears most, how many airframes of a type players have caught, what flag a ship sails under and who built her, which airline flies the most aircraft we have caught. A question only appears when there is data to answer it, so a type nobody has caught does not get a question about catches.
The same text is on the page, in the HTML a crawler reads without running JavaScript, and in the structured data. Three copies that say the same thing are easier for a search engine to trust than one.
Links into the lists
Subject pages now point into the Directory lists they belong to. A type page links to every caught airframe of that type and to the other types from the same builder. An airline page links to its caught airframes. A ship page links to the other caught ships under the same flag, from the same yard and in the same category. The breadcrumbs go through the Directory too: Skydex › Directory › Aircraft types › Boeing › 737-800.
Data you can take
The reference lists are published as JSON at /api/public/directory/export: aircraft types with model, builder, class, engines, rarity tier and how many airframes the network has heard; airlines; manufacturers; flag states; shipyards. Every row carries the URL of its page. Registrations, ICAO24 addresses and MMSI numbers are left out on purpose, because a list of individual tails with dates is a way to follow a person.
The rarity tiers themselves are already open under CC BY 4.0 in the taxonomy repository. If you use either, a link back to the page you took it from is all we ask.
For assistants
There is now a /llms.txt at the root of the site: a plain summary of what Skydex is, where the Directory sections live, how page addresses are built and what our counts mean. "Sightings", for example, means one airframe on one UTC day, never a flight. The AI crawlers we know by name are allowed in robots.txt, including the export.
Next
Search. One box across both lanes that takes a registration, a callsign, an MMSI or a type name.