ModelEndpoint API

Configure an OpenAI-compatible model backend for an Agent or Run.

Kubernetes resource

The ModelEndpoint v1alpha1 chart schema supports url or serviceDns, model, provider, and capabilities.chat, embeddings, and rerank. The Agent and Run can reference a ModelEndpoint by name and optionally override modelName.

apiVersion: core.kaigents.io/v1alpha1
kind: ModelEndpoint
metadata:
  name: local-model
  namespace: kaigents
spec:
  url: http://model-server.kaigents.svc.cluster.local:8000/v1
  model: example-model
  provider: local
  capabilities:
    chat: true

Integration requirements

Replace the sample URL and model with your actual serving endpoint; a ModelEndpoint object does not start a model server.
Check API compatibility, response limits, context window, and network reachability from the runner.
Store any required API credentials separately from the ModelEndpoint manifest.
Benchmark local and remote deployments using your quality, latency, and total-cost requirements.

Inspect

kubectl get modelendpoint local-model -n kaigents -o yaml
kubectl explain modelendpoint.spec --api-version=core.kaigents.io/v1alpha1

All documentation Getting started