Menu

Codex icon

Codex

Codex

Codex is OpenAI’s coding agent for repository work. It can inspect files, make changes, run development commands, and review code. Its interfaces include a desktop app, CLI, IDE extension, and cloud workflows, with availability depending on the account and plan. Official CLI guide.

Work with an existing repository

Install the chosen interface from the official guide, sign in, and open the project directory. Begin with a focused task such as explaining a failing test or changing one behavior. Provide the relevant files and verification command, then inspect the resulting diff and test output.

The CLI supports interactive work and repeatable codex exec workflows. Permission settings control file edits and command execution. Local tasks still need the project’s runtime and dependencies. The agent’s summary does not replace checking the application.

Free access and paid usage

The current pricing documentation lists limited desktop access on ChatGPT Free, subject to rollout. Do not assume that this includes every CLI, IDE, cloud feature, or model. Check the available interfaces and usage dashboard for your account.

Eligible paid plans include usage under their plan conditions. Codex and ChatGPT Work share usage. Additional credits can add cost. API-key sessions are billed at API rates and have different access conditions, including the absence of subscription cloud features. API prices do not predict how many tasks a subscription includes.

The CLI source uses the Apache 2.0 license. That software license does not make OpenAI model calls or cloud execution unlimited or free.

Official documentation was checked on October 23, 2026, alongside the locally installed CLI’s version and help output. No coding-quality benchmark was performed.

Quick Info

Codex icon
Category
Code Generation
Published on
Documentation reviewed
Pricing
Free Limited desktop access on Free, subject to rollout and account limits
Subscription Eligible ChatGPT plans with included usage limits
API API-key use billed separately at API rates