Implementation:Openai Openai node ChatCompletionCreateParams

Knowledge Sources	openai-node Chat Completions API
Domains	NLP, API_Design
Last Updated	2026-02-15 00:00 GMT

Overview

Concrete type definition for constructing Chat Completion request bodies provided by the openai-node SDK.

Description

ChatCompletionCreateParams is a TypeScript type (union of streaming and non-streaming variants) that defines the shape of request bodies for the /chat/completions endpoint. It enforces required fields (messages, model) and provides type-safe optional parameters for controlling generation behavior. The type is a discriminated union keyed on the stream property.

Usage

Use this type when constructing the request body for client.chat.completions.create(), client.chat.completions.stream(), or client.chat.completions.parse(). TypeScript will enforce correct parameter shapes at compile time.

Code Reference

Source Location

• Repository: openai-node
• File: src/resources/chat/completions/completions.ts
• Lines: L1484-1950 (type union and base interface)

Signature

export type ChatCompletionCreateParams =
  | ChatCompletionCreateParamsNonStreaming
  | ChatCompletionCreateParamsStreaming;

export interface ChatCompletionCreateParamsBase {
  messages: Array<ChatCompletionMessageParam>;
  model: string | ChatModel;
  stream?: boolean | null;
  temperature?: number | null;
  max_completion_tokens?: number | null;
  top_p?: number | null;
  n?: number | null;
  stop?: string | null | Array<string>;
  presence_penalty?: number | null;
  frequency_penalty?: number | null;
  tools?: Array<ChatCompletionTool>;
  tool_choice?: ChatCompletionToolChoiceOption;
  response_format?: ResponseFormatText | ResponseFormatJSONObject | ResponseFormatJSONSchema;
  seed?: number | null;
  logprobs?: boolean | null;
  // ... additional optional parameters
}

Import

import { type ChatCompletionCreateParams } from 'openai/resources/chat/completions';

I/O Contract

Inputs

Name	Type	Required	Description
messages	Array<ChatCompletionMessageParam>	Yes	Conversation messages with role and content
model	ChatModel	Yes	Model ID (e.g., 'gpt-4o', 'gpt-4o-mini')
stream	boolean	No	Enable streaming response chunks
temperature	number	No	Sampling temperature (0-2, default 1)
max_completion_tokens	number	No	Maximum tokens to generate
tools	Array<ChatCompletionTool>	No	Tool/function definitions for function calling
response_format	ResponseFormat	No	Output format (text, json_object, json_schema)
tool_choice	ChatCompletionToolChoiceOption	No	Tool selection strategy ('none', 'auto', 'required')

Outputs

Name	Type	Description
(type only)	ChatCompletionCreateParams	This is a TypeScript type, not a runtime function. It produces a typed request object used as input to Completions.create().

Usage Examples

Non-Streaming Request

import OpenAI from 'openai';

const client = new OpenAI();

const params: OpenAI.ChatCompletionCreateParams = {
  model: 'gpt-4o',
  messages: [
    { role: 'system', content: 'You are a helpful assistant.' },
    { role: 'user', content: 'Hello!' },
  ],
  temperature: 0.7,
  max_completion_tokens: 256,
};

const completion = await client.chat.completions.create(params);

Streaming Request

const stream = await client.chat.completions.create({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'Tell me a story.' }],
  stream: true,
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content || '');
}

Related Pages

Implements Principle

• Principle:Openai_Openai_node_Request_Construction