Skip to content
← Back

How this works — and what it rests on.

This page is the long answer. It exists so you never have to take the short one on trust.

AreWeMe turns real conversations into written readings. You answer questions about what you believe — it connects what you each said with relationship research, and writes what it finds.

Understanding yourself is already work. Understanding two people — and what happens between them — is more than double the work. So we do the reading-up, and you do the understanding.

What goes in

355 questions across 17 topics. Both of you answer the same questions — each giving your own answers.

What happens

Deterministic scoring — the same answers always produce the same scores. Your pair is then read against named couple patterns from relationship research, each with its evidence shown. A language model drafts the wording of your reading; the reasoning is computed algorithmically first.

Where your answers live

They train nothing — no model learns from them. They stay for as long as your account exists, because they are how the app sees you. No human at AreWeMe reads them. Close your account, and they go with it.

Most of what well-known frameworks measure is behaviour — and behaviour grows out of beliefs and values. In testing, the engine worked out from those roots alone what several widely used frameworks would say about a person, without their questionnaires. That check is why we say these questions sit at the root.

What it rests on

The questions come from how philosophers and thinkers have answered them for centuries. What we notice about the two of you comes from relationship research — and we tell you how strongly each thing is supported.

Every pattern in a reading carries its evidence tier: strongly supported · moderately supported · worth exploring.

The research families we draw on include attachment research, Gottman's descriptive constructs (we use his descriptions, never any prediction claim), interdependence theory, family systems, relational dialectics, attribution research and capitalisation.

Structured, escalating, mutual questions are a long-studied way to build closeness. That research is this product's ancestor.

The old iceberg picture holds here: beliefs and values sit under the waterline, and behaviour builds on top. Start at the base, and the rest gets easier to talk about.

Philosophical counselling has quietly existed for decades: getting clear on what words like loyalty, freedom or enough actually mean to you turns out to be half the work. AreWeMe brings that intent to more than one person — once you've talked the base layer through, you share a language.

Your answers land in a structure, not a prompt.

MiraDevone reading

Philosophy, sociology and psychology hold more about people than anyone could read — and years pass before any of it is usable.

AreWeMe gives you structured access — the words, the wisdom and the moves you need, through a method, not a prompt.

A chatbot answers for everyone.
This one answers for you — and won't flatter you into being right.

Built for two, from the ground up.

1One body among many — doable.
2Two bodies now — each one bends the other.
3Many times more tangled — studied for decades.
Mira

Mira brings a focus on structuring your shared life around simple pleasures and avoiding unnecessary stress — believing a peaceful, low-drama relationship is the goal and that well-being comes from simplicity.

Dev

Dev brings a belief that pursuing ambitious goals and embracing some stress together is necessary for a fulfilling relationship — that challenge makes you grow and that intensity can be energising.

What the difference creates

Dev pushes for growth and meaning through difficulty, while Mira pulls towards preservation and peace.

Real readings run longer. These are lightly cropped and the names changed — engine words kept as written.

Why now

Connection matters most — and it's exactly what we're losing.

Good relationships predict a happier, healthier life — more than money or fame ever will.

Harvard Study of Adult Development

One in six people worldwide is affected by loneliness.

World Health Organization · 2025

When people bring their feelings to AI, what they ask about most is their relationships.

Anthropic · 2025

What this is not for

Not therapy. Not diagnosis. Not prediction. Not a compatibility verdict — and never meant as evidence in any proceeding. It never says whether you should be together; it can't, and it doesn't try.

The honest limit

Your personalities explain only a small part of how a relationship goes — which is exactly why this predicts nothing. The rest is what the two of you do, and that is what the evenings are for.

Who built this

The engine is an in-house algorithm — three years of the founder's research, tested first on his own relationship: from two people who didn't know each other to a happily married couple. The app grew with the relationship.

InriaBuilt at Inria Startup Studio, Sophia Antipolis. Privacy tested and approved by Inria, the French national institute for digital research.
Start nowFree for the first couples · Private by design · English & Français