Quickstart

This guide walks through registering a model, creating a version, sending prediction logs, checking drift status, and running a fairness evaluation using the REST API.

Prerequisites

• A Kora AI Governance account (sign up)
• Your API key (starts with test_ for sandbox)
• Your Tenant ID (UUID from your dashboard)

Step 1: Register a model

Create an entry in the model registry.

cURL

curl -X POST https://api.korastratum.com/ai-governance/api/v1/models \
  -H "Authorization: Bearer $AG_API_KEY" \
  -H "X-Tenant-ID: $AG_TENANT_ID" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "credit-risk-scorer",
    "display_name": "Credit Risk Scoring Model",
    "description": "Predicts probability of default for retail loan applications",
    "use_case": "CREDIT_SCORING",
    "risk_tier": "HIGH",
    "owner_team": "risk-analytics",
    "data_sensitivity": "PII",
    "framework": "XGBoost",
    "tags": ["retail", "lending", "production"]
  }'

Node.js

const response = await fetch("https://api.korastratum.com/ai-governance/api/v1/models", {
  method: "POST",
  headers: {
    "Authorization": `Bearer ${process.env.AG_API_KEY}`,
    "X-Tenant-ID": process.env.AG_TENANT_ID,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    name: "credit-risk-scorer",
    display_name: "Credit Risk Scoring Model",
    description: "Predicts probability of default for retail loan applications",
    use_case: "CREDIT_SCORING",
    risk_tier: "HIGH",
    owner_team: "risk-analytics",
    data_sensitivity: "PII",
    framework: "XGBoost",
    tags: ["retail", "lending", "production"],
  }),
});

const model = await response.json();
console.log(model.id); // "a1b2c3d4-..."

Python

import requests
import os

response = requests.post(
    "https://api.korastratum.com/ai-governance/api/v1/models",
    headers={
        "Authorization": f"Bearer {os.environ['AG_API_KEY']}",
        "X-Tenant-ID": os.environ["AG_TENANT_ID"],
    },
    json={
        "name": "credit-risk-scorer",
        "display_name": "Credit Risk Scoring Model",
        "description": "Predicts probability of default for retail loan applications",
        "use_case": "CREDIT_SCORING",
        "risk_tier": "HIGH",
        "owner_team": "risk-analytics",
        "data_sensitivity": "PII",
        "framework": "XGBoost",
        "tags": ["retail", "lending", "production"],
    },
)

model = response.json()
print(model["id"])  # "a1b2c3d4-..."

Go

req := aigovernance.CreateModelRequest{
    Name:            "credit-risk-scorer",
    DisplayName:     "Credit Risk Scoring Model",
    Description:     "Predicts probability of default for retail loan applications",
    UseCase:         "CREDIT_SCORING",
    RiskTier:        "HIGH",
    OwnerTeam:       "risk-analytics",
    DataSensitivity: "PII",
    Framework:       "XGBoost",
    Tags:            []string{"retail", "lending", "production"},
}
model, err := client.CreateModel(ctx, req)
fmt.Println(model.ID) // "a1b2c3d4-..."

Save the returned id — you'll use it for all subsequent operations.

Step 2: Create a model version

Push a new version with training metadata and performance benchmarks.

cURL

curl -X POST https://api.korastratum.com/ai-governance/api/v1/models/$MODEL_ID/versions \
  -H "Authorization: Bearer $AG_API_KEY" \
  -H "X-Tenant-ID: $AG_TENANT_ID" \
  -H "Content-Type: application/json" \
  -d '{
    "version": "1.2.0",
    "stage": "STAGING",
    "training_dataset": "retail_loans_2024_2025",
    "training_samples": 1250000,
    "metrics": {
      "auc_roc": 0.892,
      "gini": 0.784,
      "ks_statistic": 0.612,
      "accuracy": 0.867
    },
    "artifact_uri": "gs://models/credit-risk-scorer/v1.2.0/model.xgb"
  }'

Node.js

const version = await fetch(
  `https://api.korastratum.com/ai-governance/api/v1/models/${modelId}/versions`,
  {
    method: "POST",
    headers: {
      "Authorization": `Bearer ${process.env.AG_API_KEY}`,
      "X-Tenant-ID": process.env.AG_TENANT_ID,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      version: "1.2.0",
      stage: "STAGING",
      training_dataset: "retail_loans_2024_2025",
      training_samples: 1250000,
      metrics: {
        auc_roc: 0.892,
        gini: 0.784,
        ks_statistic: 0.612,
        accuracy: 0.867,
      },
      artifact_uri: "gs://models/credit-risk-scorer/v1.2.0/model.xgb",
    }),
  }
);

const data = await version.json();
console.log(data.version_id); // "v1a2b3c4-..."

Python

version = requests.post(
    f"https://api.korastratum.com/ai-governance/api/v1/models/{model_id}/versions",
    headers={
        "Authorization": f"Bearer {os.environ['AG_API_KEY']}",
        "X-Tenant-ID": os.environ["AG_TENANT_ID"],
    },
    json={
        "version": "1.2.0",
        "stage": "STAGING",
        "training_dataset": "retail_loans_2024_2025",
        "training_samples": 1250000,
        "metrics": {
            "auc_roc": 0.892,
            "gini": 0.784,
            "ks_statistic": 0.612,
            "accuracy": 0.867,
        },
        "artifact_uri": "gs://models/credit-risk-scorer/v1.2.0/model.xgb",
    },
)

data = version.json()
print(data["version_id"])  # "v1a2b3c4-..."

Go

version, err := client.CreateModelVersion(ctx, modelID, aigovernance.CreateVersionRequest{
    Version:          "1.2.0",
    Stage:            "STAGING",
    TrainingDataset:  "retail_loans_2024_2025",
    TrainingSamples:  1250000,
    Metrics: map[string]float64{
        "auc_roc":      0.892,
        "gini":         0.784,
        "ks_statistic": 0.612,
        "accuracy":     0.867,
    },
    ArtifactURI: "gs://models/credit-risk-scorer/v1.2.0/model.xgb",
})
fmt.Println(version.VersionID) // "v1a2b3c4-..."

Step 3: Send prediction logs for monitoring

Push inference logs so the platform can track metrics and detect drift.

cURL

curl -X POST https://api.korastratum.com/ai-governance/api/v1/models/$MODEL_ID/predictions \
  -H "Authorization: Bearer $AG_API_KEY" \
  -H "X-Tenant-ID: $AG_TENANT_ID" \
  -H "Content-Type: application/json" \
  -d '{
    "version_id": "v1a2b3c4-...",
    "predictions": [
      {
        "prediction_id": "p001",
        "timestamp": "2026-04-07T14:30:00Z",
        "features": {"income": 85000, "age": 34, "employment_years": 6, "debt_ratio": 0.32},
        "prediction": 0.12,
        "label": "LOW_RISK",
        "latency_ms": 23
      }
    ]
  }'

Node.js

await fetch(
  `https://api.korastratum.com/ai-governance/api/v1/models/${modelId}/predictions`,
  {
    method: "POST",
    headers: {
      "Authorization": `Bearer ${process.env.AG_API_KEY}`,
      "X-Tenant-ID": process.env.AG_TENANT_ID,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      version_id: versionId,
      predictions: [
        {
          prediction_id: "p001",
          timestamp: "2026-04-07T14:30:00Z",
          features: { income: 85000, age: 34, employment_years: 6, debt_ratio: 0.32 },
          prediction: 0.12,
          label: "LOW_RISK",
          latency_ms: 23,
        },
      ],
    }),
  }
);

Python

requests.post(
    f"https://api.korastratum.com/ai-governance/api/v1/models/{model_id}/predictions",
    headers={
        "Authorization": f"Bearer {os.environ['AG_API_KEY']}",
        "X-Tenant-ID": os.environ["AG_TENANT_ID"],
    },
    json={
        "version_id": version_id,
        "predictions": [
            {
                "prediction_id": "p001",
                "timestamp": "2026-04-07T14:30:00Z",
                "features": {"income": 85000, "age": 34, "employment_years": 6, "debt_ratio": 0.32},
                "prediction": 0.12,
                "label": "LOW_RISK",
                "latency_ms": 23,
            },
        ],
    },
)

Go

err := client.LogPredictions(ctx, modelID, aigovernance.PredictionBatch{
    VersionID: versionID,
    Predictions: []aigovernance.Prediction{
        {
            PredictionID: "p001",
            Timestamp:    time.Now(),
            Features:     map[string]any{"income": 85000, "age": 34, "employment_years": 6, "debt_ratio": 0.32},
            Prediction:   0.12,
            Label:        "LOW_RISK",
            LatencyMs:    23,
        },
    },
})

Step 4: Check drift status

cURL

curl https://api.korastratum.com/ai-governance/api/v1/models/$MODEL_ID/drift?version_id=$VERSION_ID \
  -H "Authorization: Bearer $AG_API_KEY" \
  -H "X-Tenant-ID: $AG_TENANT_ID"

Node.js

const drift = await fetch(
  `https://api.korastratum.com/ai-governance/api/v1/models/${modelId}/drift?version_id=${versionId}`,
  {
    headers: {
      "Authorization": `Bearer ${process.env.AG_API_KEY}`,
      "X-Tenant-ID": process.env.AG_TENANT_ID,
    },
  }
);

const data = await drift.json();
console.log(data.overall_status);         // "HEALTHY"
console.log(data.data_drift.psi);         // 0.04
console.log(data.prediction_drift.psi);   // 0.02

Python

drift = requests.get(
    f"https://api.korastratum.com/ai-governance/api/v1/models/{model_id}/drift",
    headers={
        "Authorization": f"Bearer {os.environ['AG_API_KEY']}",
        "X-Tenant-ID": os.environ["AG_TENANT_ID"],
    },
    params={"version_id": version_id},
)

data = drift.json()
print(data["overall_status"])         # "HEALTHY"
print(data["data_drift"]["psi"])      # 0.04
print(data["prediction_drift"]["psi"])# 0.02

Go

drift, err := client.GetDriftStatus(ctx, modelID, aigovernance.DriftFilter{
    VersionID: versionID,
})
fmt.Println(drift.OverallStatus)       // "HEALTHY"
fmt.Println(drift.DataDrift.PSI)       // 0.04
fmt.Println(drift.PredictionDrift.PSI) // 0.02

Step 5: Run a fairness evaluation

cURL

curl -X POST https://api.korastratum.com/ai-governance/api/v1/models/$MODEL_ID/fairness/evaluate \
  -H "Authorization: Bearer $AG_API_KEY" \
  -H "X-Tenant-ID: $AG_TENANT_ID" \
  -H "Content-Type: application/json" \
  -d '{
    "version_id": "v1a2b3c4-...",
    "protected_attributes": ["gender", "age_group", "region"],
    "metrics": ["disparate_impact", "equalized_odds", "demographic_parity"],
    "favorable_outcome": "APPROVED",
    "dataset_id": "eval_set_2026_q1"
  }'

Node.js

const fairness = await fetch(
  `https://api.korastratum.com/ai-governance/api/v1/models/${modelId}/fairness/evaluate`,
  {
    method: "POST",
    headers: {
      "Authorization": `Bearer ${process.env.AG_API_KEY}`,
      "X-Tenant-ID": process.env.AG_TENANT_ID,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      version_id: versionId,
      protected_attributes: ["gender", "age_group", "region"],
      metrics: ["disparate_impact", "equalized_odds", "demographic_parity"],
      favorable_outcome: "APPROVED",
      dataset_id: "eval_set_2026_q1",
    }),
  }
);

const data = await fairness.json();
console.log(data.overall_status);                     // "PASS"
console.log(data.results.gender.disparate_impact);    // 0.87
console.log(data.results.age_group.disparate_impact); // 0.92

Python

fairness = requests.post(
    f"https://api.korastratum.com/ai-governance/api/v1/models/{model_id}/fairness/evaluate",
    headers={
        "Authorization": f"Bearer {os.environ['AG_API_KEY']}",
        "X-Tenant-ID": os.environ["AG_TENANT_ID"],
    },
    json={
        "version_id": version_id,
        "protected_attributes": ["gender", "age_group", "region"],
        "metrics": ["disparate_impact", "equalized_odds", "demographic_parity"],
        "favorable_outcome": "APPROVED",
        "dataset_id": "eval_set_2026_q1",
    },
)

data = fairness.json()
print(data["overall_status"])                      # "PASS"
print(data["results"]["gender"]["disparate_impact"])    # 0.87
print(data["results"]["age_group"]["disparate_impact"]) # 0.92

Go

result, err := client.EvaluateFairness(ctx, modelID, aigovernance.FairnessRequest{
    VersionID:            versionID,
    ProtectedAttributes:  []string{"gender", "age_group", "region"},
    Metrics:              []string{"disparate_impact", "equalized_odds", "demographic_parity"},
    FavorableOutcome:     "APPROVED",
    DatasetID:            "eval_set_2026_q1",
})
fmt.Println(result.OverallStatus)                     // "PASS"
fmt.Println(result.Results["gender"].DisparateImpact)  // 0.87

What's next

• Authentication — API key formats, environments, and rate limits
• API Reference — Explore every endpoint across all AI governance services