A scoring engine, built and run by an AI agent

Built entirely in Cowork, with no terminal involved, by design. The Input Card, scoring engine, and every output section run as plain JavaScript in your browser. Nothing you type is sent anywhere — the system collects no data because it has nowhere to send it.

Every fare shown is generated from a per-cabin base price plus a variation for stops, fare tier, and date flexibility — a planning model, not a fetched quote. Every mileage figure comes from hand-maintained round-trip ranges by cabin (Business, for example, runs roughly 115,000 to 320,000 miles plus taxes and fees) — the kind of range a knowledgeable advisor would quote from memory before checking live availability. Upgrade-friendliness scores are a general reputation judgment per carrier, not a live seat map. Nothing here is presented as a live quote. The homepage banner says so directly, and this page states which numbers are computed versus referenced.

Every candidate itinerary is scored 0 to 100 on Cost, Comfort, Upgrade Potential, Award Value, Flexibility, and Risk. Your ranked priority list, from Cheapest to Best Overall Experience, maps onto those six dimensions through a fixed weighting table: a rank-1 priority carries more weight than a rank-7 one. The final score is a weighted sum, shown on the results page as an explicit percentage per dimension, so the reasoning behind the recommendation is never reduced to a single number.

As with the other systems in this portfolio, this one states plainly what's real. It does not connect to Google Flights, ITA Matrix, seats.aero, or any airline system. Every number is either computed from your inputs in real time or drawn from a hand-maintained reference range. Running the same inputs twice produces the same result — the model is deterministic by design, not randomized.

This build ships as static HTML, CSS, and JavaScript to Netlify. No backend is required for a scoring engine that runs entirely in the browser. A GitHub repository will mirror the other sites' deployment model — push to main, Netlify deploys automatically — and becomes the foundation for Phase 2's live data functions once those are approved and built.

Live fare and award data is the largest remaining gap — everything shown today is a documented model, not a search result, and closing that gap is the entirety of Phase 2. Any further real spend, such as paid API keys, still requires Susan's explicit approval first, the same rule the other two sites operate under. There is also no saved-preference memory yet, so every session starts from the same defaults instead of recalling a prior trip's inputs.