openai

The official TypeScript library for the OpenAI API

weekly downloads11.8M

versionv6.22.0

licenseApache-2.0

categoryai

Overview

The openai package is the official TypeScript/JavaScript SDK for interacting with OpenAI's REST API from server-side environments like Node.js, Deno, and Bun. Built by OpenAI and auto-generated from their OpenAPI specification using Stainless, it provides type-safe access to all API endpoints including chat completions, assistants, image generation, audio transcription, fine-tuning, and embeddings.

With over 11 million weekly downloads, this package has become the de facto standard for integrating OpenAI's language models into JavaScript applications. It handles authentication, request/response formatting, streaming, error handling, and retries out of the box, eliminating the need to manually construct HTTP requests or maintain custom type definitions.

The SDK is TypeScript-first with complete type inference for all request parameters and response objects. Version 4.x and later support modern OpenAI features like the Assistants API with persistent threads, structured JSON outputs, function calling, vision capabilities, and real-time streaming. It's designed for production use with built-in support for proxies, custom fetch options, and environment-based configuration.

Developers use this package when building chatbots, content generation tools, code assistants, document analysis systems, or any application requiring access to GPT-4, GPT-3.5, DALL-E, Whisper, or other OpenAI models. The official nature ensures compatibility with new API features as they're released and provides confidence in long-term maintenance.

Quick Start

typescript

import OpenAI from 'openai';

const openai = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY
});

async function generateChatResponse() {
  const completion = await openai.chat.completions.create({
    model: 'gpt-4o',
    messages: [
      { role: 'system', content: 'You are a helpful coding assistant.' },
      { role: 'user', content: 'Explain async/await in JavaScript concisely.' }
    ],
    temperature: 0.7,
    max_tokens: 150
  });

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

async function streamResponse() {
  const stream = await openai.chat.completions.create({
    model: 'gpt-4o',
    messages: [{ role: 'user', content: 'Count from 1 to 5 slowly.' }],
    stream: true
  });

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

async function functionCalling() {
  const tools = [{
    type: 'function',
    function: {
      name: 'get_weather',
      description: 'Get current weather for a location',
      parameters: {
        type: 'object',
        properties: {
          location: { type: 'string' },
          unit: { type: 'string', enum: ['celsius', 'fahrenheit'] }
        },
        required: ['location']
      }
    }
  }];

  const response = await openai.chat.completions.create({
    model: 'gpt-4o',
    messages: [{ role: 'user', content: 'What\'s the weather in Paris?' }],
    tools
  });

  const toolCall = response.choices[0].message.tool_calls?.[0];
  if (toolCall) {
    console.log('Function:', toolCall.function.name);
    console.log('Arguments:', toolCall.function.arguments);
  }
}

generateChatResponse();
streamResponse();
functionCalling();

Use Cases

Chatbots and conversational AI: Build customer support bots, virtual assistants, or interactive chat interfaces using the chat completions API with GPT-4 or GPT-3.5-turbo. Supports streaming responses for real-time user experience and function calling to integrate external tools or databases.

Document analysis and Q&A systems: Use the Assistants API with file search capabilities to create persistent threads that analyze uploaded documents, answer questions about PDFs, or maintain context across multiple user sessions spanning days or weeks.

Content generation pipelines: Automate blog post creation, social media content, or marketing copy generation by chaining multiple completion requests with specific instructions and structured JSON outputs for consistent formatting.

Image and audio processing: Generate images with DALL-E, transcribe audio files with Whisper, or perform vision tasks by sending images alongside text prompts to multimodal models like GPT-4V for OCR, captioning, or visual analysis.

Custom model fine-tuning: Upload training datasets in JSONL format, create fine-tuning jobs for domain-specific models, and monitor training progress programmatically for specialized applications requiring tailored model behavior.

Pros & Cons

Pros

+Official SDK with guaranteed API compatibility and day-one support for new OpenAI features like GPT-4.1 and enhanced assistants
+TypeScript-first with complete type safety generated directly from OpenAPI spec, preventing runtime errors and providing excellent IDE autocomplete
+Comprehensive feature coverage including streaming, function calling, vision, audio, fine-tuning, and assistants with minimal boilerplate code
+Built-in error handling, automatic retries, and support for proxies and custom fetch configurations across Node.js, Deno, and Bun
+Simple authentication via environment variables and straightforward async/await API design that follows JavaScript conventions

Cons

−OpenAI-specific only, requiring separate SDKs or custom code when working with alternative LLM providers like Anthropic Claude or Cohere
−Server-side only with no browser support due to CORS restrictions and API key security concerns, limiting client-side use cases
−Breaking changes between major versions (v3 to v4) required migration effort for existing applications, particularly around response structures
−Large bundle size compared to minimal HTTP clients, adding overhead if only using simple completion endpoints without advanced features
−API costs are per-token and can accumulate quickly in production without proper rate limiting, caching, or monitoring implementations

Alternatives

@anthropic-ai/sdkv0.78.0

The official TypeScript library for the Anthropic API

5.6M / week

langchainv1.2.25

Typescript bindings for langchain

aillmai

1.6M / week

llamaindexv0.12.1

<p align="center"> <img height="100" width="100" alt="LlamaIndex logo" src="https://ts.llamaindex.ai/square.svg" /> </p> <h1 align="center">LlamaIndex.TS</h1> <h3 align="center"> Data framework for your LLM application. </h3>

aillmllama

95K / week

Install

bash

npm install openai

bash

pnpm add openai

bash

bun add openai

Overview

Quick Start

typescript

import OpenAI from 'openai';

const openai = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY
});

async function generateChatResponse() {
  const completion = await openai.chat.completions.create({
    model: 'gpt-4o',
    messages: [
      { role: 'system', content: 'You are a helpful coding assistant.' },
      { role: 'user', content: 'Explain async/await in JavaScript concisely.' }
    ],
    temperature: 0.7,
    max_tokens: 150
  });

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

async function streamResponse() {
  const stream = await openai.chat.completions.create({
    model: 'gpt-4o',
    messages: [{ role: 'user', content: 'Count from 1 to 5 slowly.' }],
    stream: true
  });

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

async function functionCalling() {
  const tools = [{
    type: 'function',
    function: {
      name: 'get_weather',
      description: 'Get current weather for a location',
      parameters: {
        type: 'object',
        properties: {
          location: { type: 'string' },
          unit: { type: 'string', enum: ['celsius', 'fahrenheit'] }
        },
        required: ['location']
      }
    }
  }];

  const response = await openai.chat.completions.create({
    model: 'gpt-4o',
    messages: [{ role: 'user', content: 'What\'s the weather in Paris?' }],
    tools
  });

  const toolCall = response.choices[0].message.tool_calls?.[0];
  if (toolCall) {
    console.log('Function:', toolCall.function.name);
    console.log('Arguments:', toolCall.function.arguments);
  }
}

generateChatResponse();
streamResponse();
functionCalling();

Use Cases

Pros & Cons

Pros

+Official SDK with guaranteed API compatibility and day-one support for new OpenAI features like GPT-4.1 and enhanced assistants
+TypeScript-first with complete type safety generated directly from OpenAPI spec, preventing runtime errors and providing excellent IDE autocomplete
+Comprehensive feature coverage including streaming, function calling, vision, audio, fine-tuning, and assistants with minimal boilerplate code
+Built-in error handling, automatic retries, and support for proxies and custom fetch configurations across Node.js, Deno, and Bun
+Simple authentication via environment variables and straightforward async/await API design that follows JavaScript conventions

Cons

−OpenAI-specific only, requiring separate SDKs or custom code when working with alternative LLM providers like Anthropic Claude or Cohere
−Server-side only with no browser support due to CORS restrictions and API key security concerns, limiting client-side use cases
−Breaking changes between major versions (v3 to v4) required migration effort for existing applications, particularly around response structures
−Large bundle size compared to minimal HTTP clients, adding overhead if only using simple completion endpoints without advanced features
−API costs are per-token and can accumulate quickly in production without proper rate limiting, caching, or monitoring implementations

openai

Overview

Quick Start

Use Cases

Pros & Cons

Pros

Cons

Alternatives

Related Content

Install

openai

Overview

Quick Start

Use Cases

Pros & Cons

Pros

Cons

Alternatives

Related Content

Install