codex-universal

Why we included this project

OpenAI publishes this Docker image as a reference for the base environment Codex runs in, so you can pull it locally and inspect the setup instead of treating the cloud as a black box. It is close to the real thing, though not identical, which makes it useful for debugging why a Codex session behaves differently from your own machine. The image reads CODEX_ENV_* variables to install specific versions of Python, Node, Rust, Go, Swift, Ruby, PHP, or Java, and it bundles common tooling like pyenv, poetry, uv, ruff, and bazelisk. If you want a reproducible container for AI-assisted coding, or plan to build a custom base image that matches your team's language and tooling preferences, this is a solid reference to start from.