OpenAI API

Quick Start

bun add openai
export OPENAI_API_KEY="sk-..."

import OpenAI from 'openai';

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

const response = await client.chat.completions.create({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'Hello!' }]
});

Current Models (2025)

• gpt-5.2: Most capable (128k context)
• gpt-4o: Fast multimodal (128k context)
• gpt-4o-mini: Cost-effective (128k context)
• gpt-4o-transcribe: Audio transcription optimized
• gpt-4o-mini-transcribe: Cost-effective transcription
• o1-preview: Advanced reasoning (128k context)
• o1-mini: Fast reasoning (128k context)

Chat Completions

const response = await client.chat.completions.create({
  model: 'gpt-4o',
  messages: [
    { role: 'system', content: 'You are a helpful assistant' },
    { role: 'user', content: 'Explain AI' }
  ],
  temperature: 0.7,
  max_tokens: 1000
});

Streaming

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

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

Function Calling

const response = await client.chat.completions.create({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'What is the weather?' }],
  tools: [{
    type: 'function',
    function: {
      name: 'getWeather',
      parameters: {
        type: 'object',
        properties: { location: { type: 'string' } },
        required: ['location']
      }
    }
  }]
});

Embeddings

const response = await client.embeddings.create({
  model: 'text-embedding-3-small',
  input: 'Your text here'
});

const embedding = response.data[0].embedding; // 1536 dimensions

Images (GPT-Image-1.5)

const image = await client.images.generate({
  model: 'gpt-image-1.5',
  prompt: 'A serene landscape',
  size: '1024x1024',
  quality: 'standard'  // or 'hd'
});

Audio

const transcription = await client.audio.transcriptions.create({
  file: fs.createReadStream('audio.mp3'),
  model: 'whisper-1'
});

const speech = await client.audio.speech.create({
  model: 'tts-1',
  voice: 'alloy',
  input: 'Hello world'
});

Top Errors

1. Invalid API Key (401): Verify OPENAI_API_KEY
2. Rate Limit (429): Implement exponential backoff
3. Model Not Found (404): Use correct model names
4. Context Length (400): Reduce input size
5. Invalid JSON: Fix function calling schemas

references/error-catalog.md

Resources

Reference Guides

• references/models-guide.md

  - Complete model comparison and selection

• references/function-calling-patterns.md

  - Function calling best practices

• references/structured-output-guide.md

  - Structured outputs with JSON Schema

• references/embeddings-guide.md

  - Text embeddings and vector search

• references/images-guide.md

  - GPT-Image-1.5 image generation

• references/audio-guide.md

  - Whisper transcription + TTS

• references/cost-optimization.md

  - Token optimization and pricing

• references/top-errors.md

  - Top 20 errors with solutions

• references/error-catalog.md

  - Complete error reference

Templates

• templates/basic-usage.ts

  - Quick start example

• templates/chat-completion-basic.ts

  - Basic chat completion

• templates/chat-completion-nodejs.ts

  - Node.js implementation

• templates/streaming-chat.ts

  - Streaming responses

• templates/streaming-fetch.ts

  - Streaming with fetch API

• templates/function-calling.ts

  - Tools and function calling

• templates/structured-output.ts

  - JSON Schema outputs

• templates/vision-gpt4o.ts

  - Vision with GPT-4o

• templates/embeddings.ts

  - Text embeddings

• templates/image-generation.ts

  - GPT-Image-1.5 generation

• templates/image-editing.ts

  - Image editing

• templates/audio-transcription.ts

  - Whisper transcription

• templates/text-to-speech.ts

  - TTS with voices

• templates/moderation.ts

  - Content moderation

• templates/rate-limit-handling.ts

  - Exponential backoff

• templates/cloudflare-worker.ts

  - Cloudflare Workers integration