SDKs¶

Talking to the platform¶

Because the platform is Kubernetes, the official Kubernetes client libraries work directly against it — no vendor SDK in between.

Language	Library
Go	`k8s.io/client-go`
Python	`kubernetes`
Java	`io.kubernetes:client-java`
JavaScript / TypeScript	`@kubernetes/client-node`

from kubernetes import client, config

config.load_kube_config()                       # or load_incluster_config()
v1 = client.CoreV1Api()

for pod in v1.list_namespaced_pod("<project>").items:
    print(pod.metadata.name, pod.status.phase)

Talking to models¶

Model endpoints and the LLM gateway are OpenAI-compatible, so the OpenAI SDK for your language works with a base URL change and a virtual key:

PythonTypeScriptcurl

from openai import OpenAI

client = OpenAI(
    base_url="https://<your-gateway-host>/v1",
    api_key="<your-virtual-key>",
)

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://<your-gateway-host>/v1",
  apiKey: process.env.GRN_API_KEY,
});

curl -sS https://<your-gateway-host>/v1/models \
  -H "Authorization: Bearer <your-virtual-key>"

Keys belong in the environment, not the source

Read keys from environment variables or a mounted Secret. A key committed to Git is a key you have to rotate.

Scaffold status

First-party SDKs for platform-specific APIs are planned to be generated from OpenAPI rather than hand-written. Until then, the Kubernetes clients above are the supported path.