Voice carries more signal than photos

Photos capture a moment. A voice captures a person — pace, warmth, humor, energy. Research consistently shows voice is one of the strongest predictors of interpersonal attraction and long-term compatibility.

That's the whole thesis behind AI voice matching.

What is a voice embedding?

When you record your voice, our AI turns it into a voice embedding — a mathematical fingerprint made up of hundreds of numbers that describe how you sound. Not what you said, but *how* you said it: tone, warmth, cadence, energy.

Two people with similar voice embeddings tend to have similar energy. Two people with genuinely complementary embeddings tend to *click*.

The compatibility score

We combine three signals to produce a real compatibility score:

Voice embedding similarity — do your voices actually fit?
Stated values — what you care about
Interests overlap — what you like doing

This isn't a "hot or not" ranking. It's an explainable score based on real signal.

Why swipe apps can't do this

Swipe apps are built around photos and text. They can re-rank a deck, but they can't match on the traits that actually matter — because they never captured them in the first place.

AI voice matching only works if the app is voice-first from day one. That's why WhatsLove was built the way it was.

The privacy piece

Voice embeddings are stored encrypted and never sold. They power matching inside WhatsLove — nothing else.

Curious what your voice fingerprint sounds like? Download WhatsLove and hear yourself matched to real chemistry.