OpenAI expands Codex beyond coding
OpenAI has announced a major Codex update that lets it operate the computer, work across more tools, generate images, and continue tasks over time.
Summary
OpenAI announced on April 16, 2026 a broad Codex update framed as an expansion far beyond traditional coding. According to the company, Codex now supports more stages of the software development lifecycle and is becoming a more comprehensive partner for developers.
Among the additions described by OpenAI are the ability to operate the user's computer alongside them, work with more tools and apps, generate images, remember preferences, learn from prior actions, and take on ongoing and repeatable work. The app also adds deeper support for pull request review, multiple files and terminals, remote devboxes over SSH, and an in-app browser.
In practice
In practice, Codex moves closer to becoming a unified workspace for development, review, and iteration. OpenAI says the in-app browser lets users comment directly on pages to provide more precise instructions to the agent. The company also says multiple agents can work in parallel on a Mac without interfering with the user's normal activity in other apps.
The update also includes more than 90 additional plugins combining skills, app integrations, and MCP servers, expanding the ways Codex can gather context and take action.
Context
OpenAI presents this update as a response to how Codex usage has evolved. According to the company, developers increasingly use it not only to write code, but also to understand systems, gather context, review work, debug issues, and keep longer-running tasks moving. The product is therefore being extended across more stages of technical work inside the same interface.
The company also says automations have been expanded to reuse existing conversations and preserve context, and that Codex can now schedule future work and resume tasks over days or weeks. Memory is being released in preview.
Why it matters
- It marks a shift in Codex positioning, from a coding tool to a broader working environment.
- It combines coding, browsing, image generation, automations, and memory in one workflow.
- It reinforces the trend toward AI products that handle recurring work rather than only one-off prompts.
- It signals growing competition around the operational layer of developer work.