How it works

A person and a sport, measured in the same units

SportSync doesn't sort you into a bucket. It places you — and every sport — as points in one behavioral space, and reads the distance between them.

The core idea

One shared coordinate space

Most "what sport should I do" tools map a personality type to a hand-written list of sports. SportSync works the other way around. You are encoded as a vector of behavioral scores. Every sport is encoded as a vector using the same scores — how much adrenaline it demands, how solo or social it is, how much it rewards routine versus variety, and so on.

A match is then just distance: which sport's point sits closest to yours. Nothing is looked up in a table. This is the engine running behind every result on the site.

You shared behavioral space you Every sport
Same axes on both sides. The match is the nearest point, with the reasons it's near.

The dimensions

Seven behavioral axes

Your answers resolve into scores along these axes. Each sport carries a value on the same axes, so the two are directly comparable.

Input

From answers to a vector

A short behavioral quiz (15–20 questions, no account) is scored into the axis values above — not by counting "right" answers, but by weighting each response's pull on each axis. That produces a first read, and one of four athletic personality types.

If you connect a wearable, two weeks of real movement data — how you actually train, recover and move — sharpens the same vector. Self-reported answers are a starting point; behavior is the correction. Quiz-only lands around 70% match confidence; with wearable data it moves past 90%.

Output, made real

The place graph

A recommendation you can't act on is trivia. Once the sport is named, SportSync connects it to real places to play — vetted venues and community groups, with schedules, contacts and locations.

Venues mapped, by city

403 Riyadh 25 Jeddah area 24 Dammam 20 Medina 18 Khobar 17 Hofuf 16 Abha 12 Qatif 12 Al Ahsa + more cities

590+ venues and 200+ community groups across 20+ Saudi cities, covering 86 sports. Expanding.

On cost and honesty

Where the language model is actually used

The vector, the scoring and the matching are structured computation — not a chatbot. The language model is used where judgment genuinely helps: running the coach's discovery conversation, and writing the reasoning behind a specific pick. It is not the compute engine for every interaction, which is what keeps the system fast and cheap to run as it scales.

One consequence worth stating plainly: SportSync names sports from the model's broad knowledge of what exists, applied through the axis method — it is not reading from a fixed catalogue of a few thousand rows. The method generalizes; the list doesn't have to be finite.

See it run on you

The quiz is free and takes a couple of minutes. Or, if you're building something that could use this layer, start here.

Take the quiz → Partner with us →