OpenAI Cookbook

The OpenAI Cookbook is a high-value implementation reference for OpenAI API usage patterns, agent workflows, evaluations, RAG, multimodal systems, Realtime/voice, fine-tuning, Codex, and production integrations.

Why It Matters

  • EXTRACTED: The current ingest discovered 235 Cookbook article/example pages from https://developers.openai.com/cookbook.
  • EXTRACTED: Every discovered developers page mapped to a source file in the MIT-licensed openai/openai-cookbook GitHub repository during this crawl.
  • INFERRED: This source should be treated as a living implementation library rather than a static tutorial because new examples and articles can appear over time.

How To Use This Hub

  • Start with openai-cookbook-taxonomy when choosing examples by domain.
  • Use individual mirrored source pages under wiki/sources/openai-cookbook/ for exact example content.
  • Check current official OpenAI docs before copying model names, SDK calls, or parameters into production code.
  • Run scripts/ingest-openai-cookbook.ps1 to refresh the mirror and detect new or changed examples.

Major Categories

  • Agents SDK / agent workflows: 16 pages
  • ChatGPT / GPT Actions: 34 pages
  • Codex / coding agents: 9 pages
  • Deep Research / MCP: 2 pages
  • Evaluation / eval flywheels: 23 pages
  • Fine-tuning / reinforcement fine-tuning: 10 pages
  • General OpenAI API patterns: 26 pages
  • GPT-5 / reasoning / prompting: 13 pages
  • gpt-oss / open-weight deployment: 11 pages
  • Multimodal / image / video: 13 pages
  • RAG / retrieval / vector databases: 46 pages
  • Realtime / voice / transcription: 14 pages
  • Responses API / tool orchestration: 3 pages
  • Structured outputs / function calling: 6 pages
  • Third-party integrations: 9 pages

Counterpoints And Gaps

  • Some Cookbook examples may lag the newest OpenAI API guidance; always check current official docs before production use.
  • The taxonomy is path-and-keyword based, so cross-cutting examples may belong to more than one category.