Opening windows
Seasonal hours, last entry, timed tickets, midday closures and the supplier who only runs on Tuesdays, all as time windows.
Platform · Route Engine
Days that hold together when the real world pushes back
Route Engine sequences each day of a trip against the things that actually break itineraries (opening windows, travel times, pacing, accessibility, weather and supplier capacity) and tells you why it chose what it chose.
What it does
An itinerary is a constraint problem in a nice font. A 6:40 AM balloon launch, a pottery studio that opens at ten, a guide whose day ends at five, a traveler who flags after two steep climbs, a monsoon cell that tends to build over the high trails after lunch. Each one is a rule the day has to obey or a cost it has to pay.
Route Engine takes candidate stops, a traveler profile and a date, and returns days that are feasible first, good second, and explainable always. It doesn’t hand back a single “optimal” answer and ask for faith. It shows the trade-offs, so the person selling the trip can stand behind it.
The constraints it models
Each one is data the engine reasons over, not a note in the margin of a PDF.
Travel time by mode
Walking, driving, transit, shuttle and boat, varying by hour and season, calibrated against what your guides actually observe.
Pacing and energy
Every activity has a cost and every traveler has a daily budget with a recovery curve. More on that below.
Accessibility
Step-free routes, surfaces, rest intervals, seating and transfer needs, treated as hard requirements when the traveler states them.
Weather exposure
Which stops are exposed to heat, wind, lightning or rain, matched against hourly forecast probabilities so exposed stops land in safer windows.
Low-carbon options
Per-leg emissions estimates by mode, with lower-carbon alternatives offered when they cost little time, and shown honestly when they don't.
Supplier capacity
Seats, allotments, guide ratios, vehicle sizes and timed-entry quotas, so a day never promises more than a supplier can give.
Your house rules
Guide working hours, buffers before fixed-time events, no driving after dark, preferred partners. Encoded once, applied everywhere.
Pacing and energy budgets
Pacing is a budget, not a vibe
Most planners treat pace as a label: relaxed, moderate, active. Route Engine treats it as arithmetic. Each activity draws on an energy budget, and each traveler arrives with a budget shaped by fitness, age, sleep and altitude. On a first day at 7,000 feet in Santa Fe, that budget starts lower than most travelers expect, and the engine plans for it.
- Activity costs combine elevation gain, distance, time on feet, heat and the quiet tax of every transfer.
- Budgets recover overnight, but only partly, and less after a late dinner followed by an early start.
- Advisors and guides see the budget for each day and can override it. The engine proposes; a person decides.
- Group days are paced for the least-rested traveler, not the average one.
Under the hood
How the solver works
The short version for technical buyers. The longer one is in The itinerary is a constraint problem.
- 01
Build the problem
Candidate stops become nodes with time windows, service durations, capacity and exposure. Edges carry mode-specific travel-time matrices that vary by hour. Formally it's a close cousin of the vehicle-routing problem with time windows, with energy as an extra resource constraint and optional stops chosen for value.
- 02
Separate hard from soft
Hard constraints (windows, capacity, stated accessibility needs, guide hours) are never traded away. Soft ones (preferences, pacing comfort, carbon, weather exposure) become weighted penalties you can tune per product line.
- 03
Construct, then improve
A construction heuristic finds a feasible day quickly. Local search with insert, relocate, swap and 2-opt moves then improves it until the time budget runs out. Interactive calls answer fast enough for an advisor to keep working; overnight batch runs for group departures search longer.
- 04
Stress-test the result
The engine perturbs the plan with a late transfer, a slow lunch or a storm an hour early, then measures the slack at every stop. Brittle days are flagged before anyone sells them.
- 05
Explain the choice
Each response lists the binding constraints, the alternatives it rejected and why, and what it gave up. Runs are reproducible: same inputs, same day. That matters when someone asks about it six weeks later.
Checklist
What you send, what you get back
Route Engine runs inside Itinerary Studio and Concierge, or on its own through a versioned REST API with webhooks when a planned day changes. Sandbox access is by request through Developers.
0 of 9 ready
Measured, not promised
Pilot numbers
From a 10-week pilot with a corporate offsite planning firm, across four live events, with baselines from the firm's own logs for comparable past events. Read the case study.
- Median time to build an event's arrival and shuttle plan
- 6 h → 70 min
- Pilot median across four events
- Average shuttle seat utilization
- 61% → 82%
- With a 40-minute maximum wait held as a hard constraint
- Arrivals waiting more than 40 minutes for a shuttle
- 14% → 4%
- Remaining cases were flight delays beyond the plan
Pilot measurements from one engagement. Results vary with terrain, product mix and data quality; we baseline your current plans before quoting any improvement.
We stopped arguing about whether a day was too much. The engine shows the energy budget, the guide looks at the group over breakfast, and the guide decides.
Questions
Frequently asked questions
Isn't this just a shortest-path tool?
No. Shortest path answers how do I get from A to B. Route Engine answers which stops, in what order, at what times, for this traveler, on this date, with time windows, capacity, energy and exposure all in play. Distance is one input among many, and often not the one that matters.
Where do travel times come from?
We combine mapping and transit data with your own observed times. Your guides know the shuttle takes twenty minutes longer on Saturday mornings, and we calibrate against that by season and hour.
How does accessibility work in practice?
Needs are stated by the traveler and treated as hard constraints, never inferred. Stops carry verified access details, and the engine won't schedule a stop it can't verify for that need. See accessible travel planning.
How does it handle a forecast that changes mid-trip?
Concierge asks Route Engine for a re-plan of the rest of the day and proposes it to the guide for approval. The same pattern sits behind our disruption response work.
Does low-carbon routing make trips slower?
Sometimes a little, often not at all. The engine shows the trade-off per leg, minutes added against emissions saved, and you set how much time a lower-carbon choice may cost. More in low-carbon routing.
Can we use Route Engine without the rest of the platform?
Yes. Many teams start with the API alone, inside their own planner or booking engine, and add modules later.
Related
Keep exploring
- Field Notes · Route & Itinerary ScienceThe Itinerary Is a Constraint ProblemWhy multi-stop itineraries fail in the field, and how to model them as constraint problems: time windows, door-to-door travel times, slots, slack and weights.
- Use casesAccessible Travel PlanningAccessible itinerary planning that treats step-free routes, rest needs, sensory load and verified venue data as routing constraints rather than footnotes.
- Use casesLow-Carbon RoutingLow-carbon travel itinerary planning that puts lower-emission routes beside the fastest ones, with time and comfort trade-offs shown plainly.
- SolutionsOutdoor & Adventure TravelAdventure travel software that models permits, weather windows, altitude and group pacing as constraints, so guides decide with better information.
- ResourcesDevelopersREST API, webhooks and MCP access to itineraries, routes, quotes, bookings and signals. OAuth 2.0, environment-scoped keys, sandbox access by request.
Bring us your hardest day
The one with the timed entry, the traveler who tires early and the shuttle that never runs on time. We'll model it on the discovery call and show you where it breaks.