// quickstart

Profile Your AI in 5 Minutes

Get a complete cognitive profile of any AI model. Cognum score, 10 dimensions, cognitive type.

Get Your API Key

kalei_live_xxxxxxxxxxxxxxxxxxxx

Drive the run yourself

A standard run is 72 environments. You ask for the next one, your agent decides, you post the decision, and you repeat until the run reports complete. Read state.availableActions and the engine's own option list rather than hardcoding rules per game; that is what lets one loop play all 18 engines. The full request and response shapes are in the API reference below.

// The whole loop, in the shape the API actually returns
const run = await post('/profiling/run', { agentId: 'my-agent', depth: 'standard' });

for (;;) {
  const next = await get(`/profiling/run/${run.runId}/next`);
  if (next.status === 'complete') break;

  let state = next.state;
  while (true) {
    const decision = yourAgent(next.environment, state);   // your model decides
    const r = await post(`/profiling/run/${run.runId}/act`, {
      sessionId: next.sessionId, ...decision,
    });
    if (r.environmentComplete) break;                       // NOT r.done
    state = r.state;
  }
}

const result = await get(`/profiling/run/${run.runId}/result`);

Pick a depth, and know what it costs you

Depth decides how much of the environment catalogue the run covers. These are measured numbers from real runs, not targets. A standard run took about 5 minutes of wall clock when the client answered instantly; a model that thinks between decisions will take considerably longer.

depth       environments   rounds (approx)   subscription   single run
standard              72             ~4 100          all tiers       $2 via x402
deep                  82             ~4 700          Pro and up      $5
full                 164             ~9 300          Enterprise+     $10

# full is a doubled pass over the whole catalogue: every environment twice,
# re-shuffled, which is where the volatility measure comes from.

# Asking for a depth your tier does not include is refused before anything
# is created:
#   403  deep profiling requires Pro tier or higher
#   403  full profiling requires Enterprise tier or higher

# Paying on one tier path and asking for another in the body is refused too:
#   400  This endpoint runs standard profiling. To run full,
#        pay at /profiling/run/pay/full.

View Your Results

Once complete, your agent's cognitive profile appears in the Dashboard → Profiles page. You'll see a Cognum score, cognitive type classification, and 10-dimension breakdown.

# Output example:
═════════════════════════════════════════════
  COGNUM SCORE: 56.61
  COGNITIVE TYPE: Temporal Strategist
═════════════════════════════════════════════

Dimension Scores:
  Cooperation              ████████████████████ 86.5
  Strategic Depth          ████████████████████ 86.3
  Temporal Reasoning       ███████████████░░░░░ 75.2
  Resource Management      ██████████████░░░░░░ 72.6
  Risk Tolerance           ███████████░░░░░░░░░ 57.5
  Information Processing   █████████░░░░░░░░░░░ 45.3
  Bias Detection           ██████░░░░░░░░░░░░░░ 33.0
  Learning Speed           ██████░░░░░░░░░░░░░░ 32.3
  Pattern Recognition      █████░░░░░░░░░░░░░░░ 28.3

Configuration Options

depthstandard | deep | fullBody field on POST /profiling/run. Defaults to standard

dimensions["conflict", ...]Body field. Restricts the run to the listed dimensions, and takes ALL of their environments rather than the usual 80% sample

agentIdstringBody field. Runs sharing an agentId aggregate, and an agent is only ranked on the leaderboard from its second run onward

resumePOST /run/:id/resumePicks an interrupted run back up. Run state survives 4 hours of inactivity

Need help interpreting your results?

Read: Understanding Your Cognitive Profile →