How it works
Four things, in this order.
The order is the whole product. Most apps start with content and hope it fits you. Arbor starts with you and derives the content.
It asks first.
Before a single word of vocabulary, Arbor asks what you do, why this language, and who you want to talk to. Not to personalise a greeting — to decide what you get taught. A nurse and a wine importer need different first hundred words, and every other app gives them the same ones.
It writes the curriculum around that.
Your answers become a situation goal and a set of life clusters, and the lessons are built from those: the sentences you would actually say, in the situations you actually face. The grammar sequence underneath is hand-written and fixed — that part is not improvised, because a curriculum that reorders itself teaches nothing reliably.
You speak out loud, every day.
The tutor remembers what you are learning this for, last week, and every word you reached for in English mid-sentence. Those gaps are the most valuable thing in the conversation, so they are captured rather than politely ignored.
¿Cómo fue tu turno anoche?
Fue largo. Tuve seis pacientes.
Seis es mucho. ¿Pudiste descansar?
Un poco. Dormí en el coche.
¿En el coche? Cuéntame.
Nothing you learn gets quietly lost.
Every word you meet becomes a card, scheduled by a forgetting curve run per word — so it comes back the day before you would have lost it, rather than on a fixed timetable that is wrong for almost everything.
And then it tells you the truth about it.
Arbor will not award you a level for finishing lessons. It infers one from four things it can observe, and where it hasn't seen enough it says so. That is the least commercially convenient decision in the product and the one we would defend hardest.
Get early access