---
title: 'POST /chat/completions'
description: 'Create a chat completion'
---

# POST /chat/completions

Create a chat completion

| Detail | Value |
|---|---|
| Operation ID | `ai.createChatCompletion` |
| Method | `POST` |
| Path | `/chat/completions` |
| Connector | [OpenAI](./) |

## Parameters

_No parameters._

## Request Body

| Field | Type |
|---|---|
| `model` | string |
| `messages` | array |
| `temperature` | number |


## Using this endpoint in a workflow

1. Add an **API Call** node to your workflow.
2. Pick your OpenAI connection from the Connection dropdown.
3. In the Operation dropdown, select `ai.createChatCompletion`.
4. Fill in the parameter fields that appear. Use `{{...}}` to reference upstream values.

**Required fields are marked — fill them or the call will fail at runtime.**

## What it returns

The `API Call` node writes the response to the workflow context:

```
{
  status: 200,
  success: true,
  data: { ...response body from the API... },
  latencyMs: 142
}
```

Reference response fields downstream as <code v-pre>{{nodeId.data.path}}</code>. The exact response shape is documented on the upstream OpenAI API reference.