Know your data sources.

US transportation infrastructure — National Transportation Atlas Database. No auth. - searchRailLines (~302K) ⭐ Class I owners (BNSF/UP/CSXT/NS/CN/CPKC), passenger flag - searchPorts (175) ⭐ vessel calls + gross tonnage by coast region - searchLocksAndDams (~196 lock-bearing default; ~92K all dams) ⭐ NID - searchAirports (~19.4K FAA) airports/heliports/seaplane bases - searchIntermodalTerminals (pipeline 1.4K / liquid bulk 640 / rail TOFC 241 / RoRo 84) - getServiceInfo upstream quirks: dams layer uses full state names not codes; pipeline terminals use Y/N/NA on CRUDE_OIL/REFINED; liquid bulk uses 'Y' and most attribute columns are sparse; rail uses 3-digit CNTYFIPS suffix without state prefix.

U.S. Customs and Border Protection — live land-border crossing wait times for the ~82 U.S. land ports of entry on the Canadian and Mexican borders. Per crossing, the operational status / delay minutes / open-lane count for commercial-vehicle, passenger-vehicle, and pedestrian lanes, broken down by sub-lane (standard, FAST, NEXUS/SENTRI, ready). Useful for cross-border logistics, supply-chain / freight timing, and field-crew border planning. Upstream is a SINGLE bulk endpoint with no server-side filtering — this connector fetches it once (cached ~10 min) and filters/shapes the result in-process. No auth.