A travel decision framework that shows its work.

You provide the trip: origin, destination, dates, cabin, budget, loyalty balances, and a ranked list of what matters most, whether that's lowest cost, most comfort, the best shot at an upgrade, the best use of miles, or some balance of those. The engine scores every candidate itinerary on seven dimensions: Cost, Comfort, Upgrade Potential, Award Value, Flexibility, Risk, and Time. Each is weighted by your ranking. What comes back is the reasoning, with the price attached to it.

No single dimension decides the outcome. An itinerary that costs less but scores lower on your top-ranked priorities will not outrank a better-balanced one, and every result shows the exact weight each dimension received, so the trade-off stays visible.

The trip doesn't have to be a plain there-and-back. Two optional fields on the Trip card change its shape. Via routes the search through a stopover city; it starts blank and its nights field defaults to 0, which connects through that city the same day rather than overnighting, because most searches don't want a stopover at all. Return from a different city is an open-jaw: fly into one city, home from another, which is what you want when the middle of the trip happens overland. Either one turns the live fare search into a real multi-city query rather than two separate one-ways, and they compose, so a stopover on the way out and a different city home can be priced as one itinerary. The open-jaw carries into a watched trip too, so the every-other-day re-check prices the same shape you searched; a stopover doesn't, and watching one of those watches the plain route instead.

Plain web pages, running the same scoring engine in your browser every time, so the same trip and the same fares always produce the same answer. Five small pieces of server code handle what a browser can't do alone. One searches live fares (/api/fares). A second saves a route you ask it to watch, protected by a shared edit key so only you can add or remove one, and a third re-checks those watched routes every other day. The last two store fares you scan yourself from Kayak or Booking, and expose a privacy-scoped read of your watched trips (destination, coordinates, and dates only) for Vinyl Scout's Concert Radar.

Trip details reach the fare search only when you run one. Cabin and traveler count go with them. Your saved wallet, the cards and balances, never leaves your browser. Mileage and award figures are hand-maintained, distance-scaled planning ranges, because award availability isn't searched yet, and the screen labels them that way wherever they appear.

The one gap left is live award-space search, seats.aero style. Until that ships, no mileage figure here is a bookable quote.

Built in phases, matching the rest of the portfolio. Every phase on the roadmap is live, from the scoring framework and the full page set through the live fare backend, the custom domain, the date-range sweep, the fare watcher, the Concert Radar cross-check, and exact-match city fields. Nothing on it is pending. Award-space search is the one thing missing, and it isn't a phase yet, because it needs an availability source this project doesn't have. The full roadmap and the guide spell out what each shipped phase actually computes.

• Will not present a reference-range price or mileage figure as a live quote. Every number on this site is labeled for what it is.
• Miles get recommended only when the computed value clears a genuinely strong bar, about 1.6 cents per mile. The Award Strategy section shows that math on every run.
• Will not optimize for a single variable. A lower price doesn't win if it costs too much on the dimensions you ranked higher.
• No accounts, no passwords, no analytics, no third-party tracking. Your saved wallet never leaves your device.