How a roster is built

Four deterministic steps, all on public FAA data.

1 · AD effectivity

We pull the AD final rule from the Federal Register API and extract its (c) Applicability paragraph — the legal statement of which make/model/serials are affected.

2 · Parse the effectivity free-text

We structure it into per-model serial windows (S/N 34-7250001 through 34-7450220), serial-by-Service-Bulletin references (serials as identified in Piper SB 1413), and EXCEPT carve-outs. Serial-by-reference ADs are flagged, never asserted.

3 · Registry join

For each affected model we query the FAA AircraftInquiry registry for every US-registered tail — N-Number, serial, owner, state — and test each serial against the AD windows.

4 · Confidence bands

BandMeaning
HIGHModel matches and the serial is inside an inline AD serial window.
MEDIUMModel matches and the AD applies to all serials, or gates serials by an external SB we can't read (serial-conditional — confirm against that SB).
LOWModel-family match only or serial unparseable.

Tails provably outside every AD window are dropped. Over-include = a wasted call; under-include = a missed prospect — so we band, we don't guess.


FAA AD Roster · Real FAA data (Federal Register Part-39 + AircraftInquiry registry). Sales-intelligence only, not an airworthiness determination — confirm per tail with an A&P/IA. · x402 · llms.txt