Live forecast

Know where the charter fleet will be, before anyone calls to book it.

Every U.S. Part 135 aircraft, tracked daily. Every market, forecast 90 days ahead. The hidden positioning cost inside every charter quote, made visible.

Read a tripOpen the map
Loading…
Tight · fewer aircraft than normalNormalLoose · more than normalCircle size = bookable jets in the market · sparks = departures

Ask it a real question.

These three answers are computed live from this morning's forecast. Click any of them to open the same question in the Quote Desk.

Built for the people who move the aircraft.

One forecast, three ways to use it. Each page keeps what you enter in its address, so you can share a link with a trip or a base already filled in.

EveryoneThe MapFourteen markets by ninety days, red where supply is tight and green where it is loose. Click any day for a plain-English reading. BrokersQuote DeskEnter a trip. Get the pressure verdict, the return-leg read, the talking points for the operator call, and the dates that would be cheaper. OperatorsPositioning BoardMarkets ranked by how short of your class they will be, the fractional fleets alongside, and a leg check for any discounted one-way. Fractional & membershipFleet PlannerWhere the programs are massing, event build-ups and the exodus after, a peak-day calendar, and the coverage gaps that put a guarantee at risk.

How it works, in one breath.

Public data, a foundation model, and a daily routine. Nothing here is guessed.

Every nightTrack the fleetOpen ADS-B data for every aircraft on a U.S. charter certificate, reduced to where each one flew and where it slept.
ThenCount the marketsFourteen regions, seven classes, two fleets: bookable charter and fractional programs, counted every day since November 2023.
Every morningForecast 90 daysGoogle's TimesFM 3.0 reads two years of history plus the holiday and event calendar, and writes the next 90 days with ranges.
AlwaysScore itselfEach day's actual count is checked against the forecasts made 7, 30 and 90 days earlier. The report card is public in the API.

For software and AI agents.

The same forecasts as JSON, and as a Model Context Protocol server any assistant can connect to.

EndpointWhat it answers
/api/regionsRegions, classes, buckets, data vintage.
/api/forecast/supplyAircraft on the ground for a region or airport, class and date, with bands and the yearly baseline. Try it
/api/forecast/repositioningDay-by-day departures versus baseline over a range. Try it
/api/quote/pressureLow, neutral or high pressure for a trip, with the reasoning. Try it
/api/forecast/accuracyRolling error and band coverage by region, class and lead time. Try it

Classes: very_light light midsize super_mid large ultra_long turboprop · Buckets: open_charter (bookable) and fractional_captive.

{"region":"SW_FLORIDA","class":"light","date":"…",
 "on_ground":{"p10":33.7,"p50":40.8,"p90":49.4},
 "baseline_365d":{"p20":32.4,"p50":38.3,"p80":44.9},
 "made_on":"…"}

Connect an AI assistant:

claude mcp add --transport http jetfinder-forecast \
  https://forecast.thejetfinder.com/mcp

Tools: list_regions, get_supply_forecast, get_repositioning_forecast, get_quote_pressure, get_forecast_accuracy.

What it is, and what it isn't.

It forecasts supply, not price. Supply is the largest driver of positioning cost, and positioning is the largest reason two quotes for the same trip differ. Price itself comes later, once the demand and empty-leg archives have a year in them.

It gives ranges on purpose. Every number carries a 10th-to-90th percentile band, and the accuracy endpoint shows how often reality lands inside it. Use the band.

It counts what it can see. Owners can block their aircraft from public tracking; large-cabin counts run lower than reality for that reason. The system measures the gap rather than hiding it.

Data: adsb.lol open data (ODbL), FAA D085 operator list and aircraft registry, OurAirports. Model: Google Research TimesFM 3.0, zero-shot.