// Track

Getting Started with OpenAI Codex

Ship Code with OpenAI's Agent

From zero to shipping with OpenAI Codex — start with the standalone CLI and ChatGPT sign-in, learn the agent loop, and ship a verified feature with explicit approval_policy and sandbox_mode controls. Direct Responses API lessons are an optional metered-billing fork.

Recommended: Complete Basic Training first

12 lessons~109 min total

Lessons are shown in recommended order. Complete them in sequence for the best experience — or jump to any lesson.

Lesson 524·7 min read

What Is OpenAI Codex?

OpenAI Codex is an agentic coding tool that reads your repository, edits files, and runs commands — a new category beyond autocomplete or chat.

Lesson 525·8 min read

Your OpenAI Account, API Keys, and Auth

Two paths into OpenAI — subscription sign-in or API key. Know which surface uses which, how to store credentials safely, and how to never leak a key.

Lesson 526·9 min read

Installing and Running the Codex CLI

From zero to running your first agentic coding session: install the Codex CLI, authenticate, and understand the interactive session lifecycle.

Lesson 527·11 min read

The Codex Loop: How the Agent Thinks

Read Codex's observable action trace, recognize repeated-strategy thrash, and stop after two failed attempts to replan from evidence.

Lesson 528·10 min read

The Models Behind Codex

Choose a Codex model with published rate-card inputs, then benchmark quality and p95 latency on your own workload instead of trusting invented multipliers.

Lesson 529·9 min read

Prompting Codex and the AGENTS.md File

Tight task prompts and a well-maintained AGENTS.md are the two cheapest investments that produce the biggest reliability gains in Codex-powered workflows.

Lesson 530·9 min read

Approvals, Sandboxing, and Safety

Codex autonomy is set by two independent dials — an approval policy and a sandbox mode — learn the blast-radius mental model that lets you choose the right combination.

Lesson 531·8 min read

The Responses API: Your First Direct Call

Drop below the Codex agent to the raw OpenAI API — learn when to call it directly and how to make your first Responses API request.

Lesson 532·10 min read

Tool Calling and Structured Output

Connect models to real data with tool calling, and lock response shapes with structured output — the two primitives that turn a chatbot into a production API.

Lesson 533·9 min read

MCP: Extending Codex with Your Own Tools

MCP is the open standard that turns Codex from a code generator into an agent that can read your database, query your APIs, and remember context across sessions.

Lesson 534·9 min read

Codex in the Cloud and Integrations

Delegate an asynchronous Codex cloud task and return to a reviewable result or diff; connected workflows may turn that result into a pull request.

Lesson 535·10 min read

Capstone: Ship a Feature with Codex

Scope it, write the AGENTS.md, run the session, review the diff, test, PR, verify — this lesson walks the full pipeline end-to-end and defines what done actually means.