Quickstart

• A TokenAir account with API access enabled.
• A TokenAir API key from Console > API Key Management.
• A terminal with curl. Node.js 18+ or Python 3.9+ is optional for the SDK examples.

Keep credentials out of source code. Set the base URL and API key in your shell or deployment environment.

export TOKENAIR_BASE_URL="https://api.tokenair.ai/v1"
export TOKENAIR_API_KEY="your_tokenair_api_key"

Use /chat/completions to verify that your key, base URL, and selected model ID work before changing application code.

curl "$TOKENAIR_BASE_URL/chat/completions" \
  -H "Authorization: Bearer $TOKENAIR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5-mini",
    "messages": [
      { "role": "user", "content": "Reply with one short sentence." }
    ]
  }'

Install the SDK if your app runs on Node.js, then keep the existing OpenAI client shape and point baseURL at TokenAir.

npm install openai

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.TOKENAIR_API_KEY,
  baseURL: process.env.TOKENAIR_BASE_URL,
});

const response = await client.chat.completions.create({
  model: "gpt-5-mini",
  messages: [{ role: "user", content: "Hello TokenAir" }],
});

console.log(response.choices[0].message.content);

Python uses the same two environment variables and the same OpenAI-compatible chat completions shape.

pip install --upgrade openai

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["TOKENAIR_API_KEY"],
    base_url=os.environ["TOKENAIR_BASE_URL"],
)

response = client.chat.completions.create(
    model="gpt-5-mini",
    messages=[{"role": "user", "content": "Hello TokenAir"}],
)

print(response.choices[0].message.content)

A successful call returns a chat completion response with achoices array. If you receive a 401 response, check the API key. If you receive a 404 or model error, confirm that the model ID is enabled for your account.

You now have a working TokenAir request path using the same OpenAI-compatible request structure your app can use in production. Keep the key server-side, then move on to migration and error handling before routing real user traffic.