Our story
Airlines have FOQA programs. Investigators have flight data recorders. Everyone else gets a track line on a map. We thought that was strange — flight data is being broadcast in the clear, all day, every day, by every modern aircraft in the sky. We just had to do something with it.
Aviation is the most heavily-instrumented industry in the world. Every modern airliner reports its position, altitude, ground speed, vertical rate, and squawk code several times per second over a public, unencrypted radio standard called ADS-B. A global volunteer network of ground receivers picks that up and aggregates it. The data is right there.
What's been missing is everything you'd actually do with it. We take the raw trace and run it through the same kind of pipeline an airline FOQA program would — flight-phase classification, attitude reconstruction, runway-event detection, stable-approach gating, wind-corrected airspeeds, incident heuristics — and then we put the result on a page anyone can read. Pilots study their own approaches. Instructors build debriefs. Investigators sanity-check. Av geeks watch the climb out of Hong Kong from row 27 of a Cathay 777.
That's the whole idea: flight analytics belongs to anyone who's curious enough to look.
Pilots flying their own aircraft pull their flights up the next morning to check the approach: was that touchdown really 1,200 ft past the threshold? How stable was the descent? Was the crosswind worse than I remember?
Flight instructors use the replay to teach approach stability without burning Hobbs time. Researchers and journalists pull bulk data through the API to build studies on go-around rates, runway utilization, and airspace congestion.
Watching the climb-out of an A380 from Heathrow at the right pitch and roll is its own thing. Searching for "flights that went around at JFK last week" and finding seven of them is its own thing too.
SkyMeter sits on top of a continuously-running ingestion pipeline. ADS-B traces arrive from the global volunteer receiver network and pass through flight splitting, geographic enrichment, hybrid interpolation, runway detection, go-around detection, and physics-based attitude reconstruction.
Weather is bonded onto every sample by a separate service that pulls METAR observations from NOAA and IEM and combines them with NOAA GFS forecasts and ECMWF ERA5 reanalysis fields. The result is a true wind triangle at every point in the flight — which is how the cockpit panel can show real Indicated Airspeed and Mach instead of the GPS ground speed everyone else gives you.
Storage is split across high-throughput analytics tables (so the airport and runway dashboards return in under a hundred milliseconds) and an immutable Protocol Buffers archive (so the replay pages can stream the full enriched trace to your browser). Every layer is built to scale horizontally — because the most-watched flight could be one that landed at any airport in the world.
If you want to know exactly how a particular detector works — go-around, stable-approach gates, attitude reconstruction — there's an answer in the FAQ. We try to be transparent about what's measured versus what's inferred.
The dial we're turning is more accessible, not more
proprietary. The processed flight files are released under the Open Database
License — every .smr download is free and attributed. Convenience
formats (FDR for X-Plane and MSFS, Garmin-style CSV, KML, the API) are how we
pay for the servers.
On the roadmap: tighter attitude reconstruction for general-aviation aircraft, ML-driven anomaly detection for the parts of the flight envelope that don't cleanly fit a heuristic, deeper weather (turbulence, icing, convection proximity), and a proper post-flight analysis report you can share with an instructor or co-pilot in a click. We listen to pilots first — if you're a CFI, ATP, or just someone who flies regularly and has a list of things you wish a tool like this did, please write to us.
SkyMeter was started in New York City by a small team that overlaps software engineering and the cockpit. Every detector in the pipeline has been validated against real-world references — pitch and roll against Garmin AHRS exports, stable-approach gates against the Flight Safety Foundation criteria, go-around detection against the MDPI 2020 method paper, weather against METAR and ERA5 ground-truth.
We are a Delaware C-corporation, run from New York, with infrastructure spread across multiple continents. If you'd like to get in touch — about a flight that doesn't look right, a feature you'd love to see, an integration, or anything else — write to contact@skymeter.io. A real person reads every message.
Tail numbers, callsigns, airports, routes, dates — start anywhere. Most flights are searchable within 24 hours of landing.