KVision

The elite tracking data. Free.

Read from the film you already have.

The elite tier, free

The data only the richest programs had.

Installed tracking cost five and six figures and needed in-stadium hardware. Now any program gets it from film, free.

The old way: installed tracking
Five and six figures a year
In-stadium hardware
The top tier only
On the OS
Tracking-grade data, free
From the film you already have
No hardware, no contract
Every level
The granular data, from film

This is what "granular" actually means.

Every measurement below the event, the detail that says what a play really was. Here it is, sport by sport.

What KVision tracks · Basketball
A made three
The event a tagger logs, and the tracking below it.
From film
Shot type
catch-and-shoot vs off-dribble
Contest / closeout distance
4.1 ft
Release time
0.68 s
Dribbles before shot
0
Movement into the shot
off a screen
Balance at release
feet set, squared
Shot distance
24.6 ft
The data that separates a spot-up from a contested pull-up.
KVision tracks this data from film. The engine reads it into the rating.

KVision tracks every detail below the play, from film. That is the granular data the engine runs on, and the data the elite tier pays six figures for.

Synergy spent two decades building the event tags, and it still does not produce what this produces, because the value was never in the tags. It is in the architecture that turns them into an evaluation.

And now that AI vision hands detection and tracking to anybody who wants them, the tagging layer that used to be the moat is a commodity. KVision's granularity sits above it, and the evaluation sits above that.

Everyone is about to be able to see the game. Almost nobody will be able to read it.

Any film, any sport, every level

One pipeline resolves any source.

A parent's phone from the bleachers, a school stream, a broadcast feed, whatever a program already records. One pipeline resolves it, then a sport-specific layer reads what it detected. No installed hardware, nothing to set up.

Phone from the bleachers
A parent's footage is enough to start.
School streams
Whatever the program already records.
Broadcast feeds
Any standard game broadcast.
Whatever the program records
Multi-cam, ceiling, or a single angle.
Live or post-game
During the game, or after, on standard cloud.
How it reads any angle
Calibration maps any camera angle onto the court or field. One core does the seeing in every sport: it finds the players, tracks the ball, keeps everyone identified, and reads the action. Then a per-sport layer names what it saw, the same movement is a pick-and-roll in basketball and a defensive line breaking in football. Same core, a different vocabulary per sport.
Depth of capture is depth of reasoning

The market's ceiling is a data ceiling.

You cannot reason past the data you capture. The market stops at the event, so it stops at shallower reasoning. KVision captures the detail below it, so the engine reasons at a level the market cannot reach, for any program on the OS, not just the ones that can pay for hardware.

The capability ladder

Five rungs. Only one of them scores the film.

The first four stop at the event, short of the dimensions a real evaluation requires. KVision is the only layer that scores them, and the only one that does it from film instead of installed hardware.

The layerWho sells itWhere it stops
1
Public data
KenPom, Opta
Stops at the event
2
Vendor event tagging
Synergy, Wyscout, PFF
Stops at the event
3
AI event tagging
The new entrants
Stops at the event
4
Tracking and wearables
Second Spectrum, Catapult
Installed hardware, and still stops short of the trait dimensions
5
KVision
KaNeXT
Scores the dimensions. From film. No hardware.

Four of those rungs are somebody's whole company. All four of them are a line item here.

What it absorbs

Every layer they sell separately is a built-in feature here.

Film and hosting
Hudl, Veo
a KTV function
Tagging and analytics
Synergy, PFF
a KVision function
Tracking and wearables
Catapult, Second Spectrum
an input tier the engine generates from film
Data feeds
Sportradar
a source the engine already pulls for free

Every one of those is a separate contract a program pays for. Here it is absorbed. A built-in feature, not a line item.

It compounds

The longer you stay, the sharper it gets about you.

Every team starts with a real read from public data on day one. Feed film, and the engine is reading your actual possessions. The longer you stay, the sharper it gets about you.

V1 · Public data75 to 84%Everyone has it. The engine evaluates anyone from public data alone.
V2 · One season of film88 to 95%Join and feed film. Confidence climbs with the dimensions only film can resolve.
V3 · Multi-season92 to 97%Stay. It compounds season over season.

Where you land inside each band depends on the sport, the film, and how much of it you feed. The engine reports a range because it has one. A single number would be tidier and less true.

V1 is what everyone has. V2 is what you get when you join. V3 is what you get when you stay.

Why the streaming is free

You do not upload to KVision. You stream on KTV.

Every game streamed free on KTV is the data event. The free broadcast pulls in the film. The film sharpens the intelligence. And the sharper intelligence pulls in more film.

The thing that looks like free streaming is how the engine gets fed.

That is not a hidden cost. It is the whole reason we can afford to give away what everybody else is selling.