> ## Documentation Index
> Fetch the complete documentation index at: https://docs.openserv.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Make your first API request

> Send your first SERV Reasoning request with the OpenAI SDK.

<Frame>
  <img src="https://mintcdn.com/openserv/wCXyJJTiOX6drWSJ/images/tutorials/first-request.webp?fit=max&auto=format&n=wCXyJJTiOX6drWSJ&q=85&s=bad637d701be76d9ffc4888c79513325" alt="A request traveling from application code through an API gateway to a language model" width="1536" height="1024" data-path="images/tutorials/first-request.webp" />
</Frame>

```bash theme={null}
npm install openai
```

Create a server-side file such as `example.mjs`. Keep the API key in an environment variable; do not expose it in browser JavaScript.

```js theme={null}
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://inference-api.openserv.ai/v1",
  apiKey: process.env.SERV_API_KEY,
});

const response = await client.chat.completions.create({
  model: "gpt-5.4-mini",
  reasoning_effort: "low",
  messages: [
    { role: "system", content: "You classify support requests." },
    { role: "user", content: "I was charged twice for the same order." },
  ],
});

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

Every request must include a system prompt, developer message, or `instructions` field. Browse the [model catalog](../models) for supported model IDs.

The response uses the standard Chat Completions shape. The generated text is in `choices[0].message.content`, and usage data is in `response.usage`.

If the request fails, check the API key, organization balance, model ID, and system prompt first. For endpoint-specific fields, see the [Chat Completions reference](../api/chat-completions).