What Vairi's matching uses, what it ignores, and what it does not promise
Vairi keeps the exact scoring model private, the same way any matching system keeps its ranking logic private. This page explains the inputs, the exclusions, and the limits in plain terms.
What goes in
Onboarding is a short AI-led conversation, not a form. It reads for behavioural signal: how someone tends to handle disagreement, how much social contact they can sustain before needing to recharge, and whether they orient toward the past, present, or future.
This signal is not a one-time snapshot: every conversation Vairi facilitates adds more read on how you actually communicate, so match quality tends to improve the longer you use Vairi, not just at signup.
What we deliberately do not use
- No photo or appearance-based ranking, no swiping, no public profile grid.
- No static personality-type quiz.
- No shared-interest matching as the primary signal.
- No follower counts or social proof as a ranking input.
The limits
Matching quality depends on the active member pool in a given city. Behavioural read from a short conversation is an inference, not a certainty. Vairi does not guarantee a specific outcome.