A model of one. And the one is you.
Every recommendation engine you have used was trained on everybody. Byway ML is pointed at a single traveler. It learns from the roads you keep and the stops you linger at, and the model it builds is yours alone.
You drive
Nothing to fill in. Arriving at a stop, saving one, rating one — every signal the model reads comes from something you actually did on the road.
The model moves
Those signals become small signed nudges across a 32-dimension vector, applied on a pass of their own. One drive tilts the model; it never overwrites it. A replayed signal is a no-op, so the same stop cannot count twice.
The next plan lands closer
The engine reads your vector when it chooses which verified stops to offer. The same roads, the same record, a different shortlist — the one you would have picked yourself.
Not likes nature.
These are the axes the model actually carries — the same list the favorites panel writes to and the engine scores candidates against. Your vector is a weight on every one of them at once, which is why two travelers asking for the same road get different stops.
- scenic
- viewpoint
- coast
- beach
- mountain
- nature
- waterfall
- historic
- cultural
- food
- local_gems
- hidden_gems
- hidden_towns
- family
- wildlife
- adventure
- art
- photography
- sunset
- sunrise
- quiet
- popular
- free
- paid
- easy
- strenuous
- pet_friendly
- accessible
- remote
- urban
- educational
- romantic
A brand-new traveler starts neutral: all thirty-two dimensions equal, no stereotype applied. Today's explicit choice still leads — what you tick before a drive outweighs what the model has learned, every time.
It never invents a place
Personalization chooses among verified places; it cannot conjure one. Every stop still traces to the public record behind it, whatever your vector says.
It learns from you, and nothing else
There is no write path to your model. It moves on route events — arrivals, saves, ratings — and on nothing else: not a prompt, not an operator, not us.
It stays yours
Your vector is scoped to the traveler who earned it, in the database itself. Byway AI can read it back to you and explain what it believes; it cannot edit it. We do not sell personal data or use it for targeted advertising.
It never flattens you
Where the model is thin, the engine falls back to the road and the record rather than guessing at you. A cold start is a quiet model, not a stereotype.
Where we are taking it.
The loop above runs today. Everything below is what a model of one should become — written here as intent, not as a shipping promise.
The same traveler wants a different road in February than in July, alone than with kids in the car, on two hours than on two days. The model will carry the context, not just the preference.
See what the model believes about you, in plain words. Correct it where it is wrong. Forget a stretch of road you would rather it did not remember.
A new traveler should not have to answer a questionnaire. One drive, honestly observed, should be enough to start — and the model should say how sure it is.
Web, the app, the car and your assistant read the same model, so the drive Byway offers in a chat window is the drive it offers on the dash.
A road trip is rarely one person's taste. We want to blend two models into a plan that neither traveler would call a compromise.
The most personal signals — where you slowed, where you turned around — should be able to shape your model without ever leaving your phone.
Drive it once. Then drive it again.
The model has nothing to say about you until you go somewhere. Plan a drive, keep the stops worth keeping, and watch the second plan come back closer than the first.
Questions about the model
Does Byway ML change which places exist?
No. It changes which verified places you are offered and in what order. The record is the record: every stop traces to its source whether the model likes it or not.
Can I see what it learned about me?
Yes — ask Byway AI about your taste profile and it reads back your strongest dimensions and the categories you picked yourself. It is a read-only view: the assistant can explain your model, never rewrite it.
What happens before it knows anything about me?
You start neutral, with all thirty-two dimensions weighted equally, and the interests you tick before a drive carry the plan. Learning seasons that over time; it never overrules the choice you just made.
Is my taste profile shared with other travelers?
No. Your model is scoped to your account by row-level security, and it is not pooled into anyone else's plan. We do not sell personal data or use it for targeted advertising — see the privacy policy.