Human intelligence · Model grade signals · Game data

The foundation model for symbiotic human+AI performance.

AI learns from what humans produce, not from how humans think, reason and behave. Skillprint captures fluid intelligence at scale from real gameplay, then gives AI platforms the tools to augment people.

No survey. No separate test. Signal from the games people already play.
500,000+people have played
1,000+games, quizzes and simulations
50+cognitive skills
5 minto your first strengths profile
The problem

AI learns from what humans produce. Not from how humans think, reason and behave.

Today’s training data is output: text, images, video, code. The missing layer is the process behind it. Decisions, adaptation, emotion, collaboration. Frontier models are reaching the limits of internet scale data, and scaling on that data alone does not guarantee fluid reasoning on new problems.

Why gameplay works
01

Play is honest

Nobody performs for a game. People stop presenting themselves and simply behave, which is the one thing a questionnaire can never reach.

02

Play asks for everything

Planning, focus, adapting, reacting, working with other people. A few minutes of play draws on the same things a working day does.

03

Play keeps going

People come back to a game they enjoy, so the picture keeps building. A profile that updates every session never goes stale the way a test does.

Products

One layer of human intelligence. Four ways in.

Amplify is the product people meet, whether they arrive on their own or through an organization. Flow puts it inside games other people built, and Signal measures what humans and AI do together. Every route feeds the same model.

Amplify

Players

Discover your strengths through play. Real games, not repetitive exercises, and a picture of how you think that builds session by session.

Discover your strengths
A Skillprint skill record: the Focus skill with its flow score, momentum and the sessions behind it
Amplify

Enterprise

Put Amplify inside the platform, app or program your people already use. They play for a few minutes; you get a continuous read on how they think, adapt and perform, without anyone stopping to be assessed.

Integrate Skillprint
The Skillprint player experience: a weekly strengths score, a playing streak and the game recommended for today
Flow

Developers

Plug the game you already built into the model. Flow reads every session, returns strengths and live difficulty tuning through one API, and pays you for the play.

Build with Flow
The Skillprint partner dashboard: strengths scored across a session, with mood and flow readings and the gameplay timeline behind them
Signal

AI Labs

Benchmark models against real human play, and measure the lift when people and AI work together rather than apart.

Explore the benchmark
The Skillprint Labs benchmark ranking frontier models against real human play
What comes back

Play a session. See something true about yourself.

A few minutes of play produces a real read: which skills carried the session, how your mood and flow moved through it, and where your profile shifted. Nothing is asked of you afterward.

A Skillprint skill record: the Focus skill with its flow score, momentum, sessions played and the progression profile behind them
Calm under pressureBadge earnedheld accuracy as speed climbed
+6 spatialSince last weekacross four sessions
01
Skill meters per sessionPattern recognition, spatial, attention and timing, scored from how the session actually went.
02
Mood and flow, read liveConfidence and flow tracked through the session, not asked about at the end.
03
Badges worth earningCalm Under Pressure, Insight Collector, seven day streaks. Recognition tied to real behavior.
Humans and AI

The question is not humans versus AI. It is what each is best at, and what they are best at together.

Models are measured against other models. Almost nothing measures whether a person actually did better with one. Skillprint runs the comparison inside real gameplay: people alone, AI alone, and people working with AI — and reports the lift.

Apart, and together

Three arms run blind inside real games: people unaided, AI alone, and people amplified by AI.

HumanAIHuman+AI

Lift, not leaderboard position

The number that matters is how much better a person performed with the model than without it.

Goal attainmentCostEfficacy

A live benchmark keeps us precise

Eleven frontier models, one neutral scorer, and statistics that survive being checked.

11 models95% CI
Contact

Let's build a future where AI amplifies people, not the other way around.

Tell us what you are building and we will show you what gameplay can measure inside it. Most conversations start with a thirty minute call and a live session.

A SkillprintTwelve of the cognitive skills and moods a Skillprint holds, from problem solving and memory through to focus and collaboration. Each one runs as a strand into the Skillprint mark at the center, where they combine into a single profile.Problem SolvingMemorySpeedAccuracyPattern RecognitionSpatial AwarenessLogicCreativityInnovateRelaxFocusCollaborate