API Reference

The Kora AI Governance API is organized around REST. It accepts JSON request bodies, returns JSON responses, and uses standard HTTP status codes.

Base URL

https://api.korastratum.com/ai-governance/api/v1

Authentication

All endpoints require two headers:

Authorization: Bearer <your-api-key>
X-Tenant-ID: <your-tenant-id>

See Authentication for details on API key formats and environments.

Endpoints

Models

Method	Endpoint	Description
POST	/models	Register a new model
GET	/models	List models with filtering
GET	/models/{id}	Retrieve a model by ID
PUT	/models/{id}	Update model metadata
DELETE	/models/{id}	Archive a model

Versions

Method	Endpoint	Description
POST	/models/{id}/versions	Create a new model version
GET	/models/{id}/versions	List versions for a model
GET	/models/{id}/versions/{versionId}	Retrieve a version by ID
PUT	/models/{id}/versions/{versionId}	Update version metadata
POST	/models/{id}/versions/{versionId}/promote	Promote version to next stage
POST	/models/{id}/versions/{versionId}/approve	Approve a version for production

Monitoring

Method	Endpoint	Description
POST	/models/{id}/predictions	Log prediction batch
GET	/models/{id}/metrics	Get monitoring metrics
GET	/models/{id}/metrics/timeseries	Get metric timeseries
POST	/models/{id}/alerts	Create a monitoring alert rule
GET	/models/{id}/alerts	List alert rules
GET	/models/{id}/alerts/history	Get alert history

Drift Detection

Method	Endpoint	Description
GET	/models/{id}/drift	Get current drift status
POST	/models/{id}/drift/evaluate	Run drift evaluation on demand
GET	/models/{id}/drift/history	Get drift evaluation history
PUT	/models/{id}/drift/config	Update drift detection thresholds

Fairness

Method	Endpoint	Description
POST	/models/{id}/fairness/evaluate	Run a fairness evaluation
GET	/models/{id}/fairness/evaluations	List fairness evaluations
GET	/models/{id}/fairness/evaluations/{evalId}	Retrieve evaluation results

Explainability

Method	Endpoint	Description
POST	/models/{id}/explain	Generate explanation for a prediction
POST	/models/{id}/explain/batch	Generate batch explanations
GET	/models/{id}/feature-importance	Get global feature importance

Experiments

Method	Endpoint	Description
POST	/experiments	Create an A/B experiment
GET	/experiments	List experiments
GET	/experiments/{id}	Retrieve an experiment by ID
PUT	/experiments/{id}	Update experiment configuration
POST	/experiments/{id}/start	Start an experiment
POST	/experiments/{id}/stop	Stop an experiment
GET	/experiments/{id}/results	Get experiment results

Health

Method	Endpoint	Description
GET	/health	Health check (no auth required)
GET	/ready	Readiness check (no auth required)

Response format

Successful responses return the resource directly or in a paginated wrapper:

{
  "id": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
  "name": "credit-risk-scorer",
  "risk_tier": "HIGH",
  "status": "ACTIVE",
  "latest_version": "1.2.0"
}

Paginated list responses:

{
  "data": [...],
  "next_cursor": "eyJpZCI6Ii4uLiJ9"
}

Error responses follow a consistent structure:

{
  "error": "Bad Request",
  "message": "name is required",
  "status": 400
}