Your stack, but with model="auto"
and every lab behind it.

# nexoauth — the whole integration
from openai import OpenAI
import os

client = OpenAI(
    base_url="https://api.nexoai.dev/v1",
    api_key=os.environ["NEXOAI_KEY"],
)

resp = client.chat.completions.create(
    model="auto",            # routed: latency + cost + reliability
    messages=[{"role": "user", "content": "hello"}],
)
print(resp.routed_to, resp.latency_ms)  # claude-sonnet-5 412ms

// @nexoai/sdk — TypeScript
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.nexoai.dev/v1",
  apiKey: process.env.NEXOAI_KEY,
});

const resp = await client.chat.completions.create({
  model: "auto",
  messages: [{ role: "user", content: "hello" }],
});
console.log(resp.routed_to, resp.latency_ms); // claude-sonnet-5 412ms

# curl — no SDK required
curl https://api.nexoai.dev/v1/chat/completions \
  -H "Authorization: Bearer $NEXOAI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "auto",
    "messages": [{"role": "user", "content": "hello"}]
  }'

# response carries the routing decision:
{"routed_to": "claude-sonnet-5", "latency_ms": 412, "cost": "$0.0018"}