OpenAI rolls out major Codex update with automation, plugins and computer use

OpenAI has rolled out a major set of updates to Codex, positioning it as a tool for “almost everything” in modern software and creative workflows. These updates, now rolling out on the Codex desktop app, significantly expand what Codex can see, connect to, and act on—bringing it closer to a true AI work companion rather than a narrow coding assistant.

One of the most notable changes is deeper tool integration. Codex now supports more than 90 plugins, allowing it to gather context and take action across the tools many teams already rely on. This includes documentation platforms, project management systems, code review tools, creative software, and deployment pipelines. Instead of switching contexts, users can increasingly orchestrate work directly through Codex, reducing friction across complex workflows.
Automation has also taken a meaningful step forward. Codex automations can now run within the same conversation thread, preserving full context as tasks progress. This allows Codex to pause, resume, and continue long‑running or multi‑step work without losing track of prior decisions. It can schedule future actions, follow up on open pull requests, and stay on top of recurring tasks—effectively acting as a persistent collaborator rather than a one‑off assistant.
Visual creation is another area seeing rapid expansion. With GPT‑Image‑1.5 integrated into Codex, users can now generate and iterate on images directly within their workflow. This makes it easier to create frontend designs, mockups, game assets, and other visual components without leaving the development environment. Importantly, this capability is included with a ChatGPT account, with no separate API key required, lowering the barrier for experimentation and iteration.
On macOS, Codex now gains “computer use” capabilities. This means it can interact with applications by seeing the screen, clicking, and typing with its own cursor. Running quietly in the background, Codex can handle tasks such as frontend iteration, application testing, and workflows that do not expose APIs. This opens up automation possibilities even for legacy tools or custom software that previously resisted integration.
Taken together, these updates signal a shift in how AI assistants fit into daily work. Codex is no longer just about writing code faster—it’s about connecting tools, maintaining context over time, learning user preferences, and taking on repeatable work across disciplines. As these capabilities mature, Codex is shaping up to be less of a helper on demand and more of a continuously available partner embedded directly into how work gets done.

Comments

Popular posts from this blog

Anthropic launches Claude Design, a new product for creating quick visuals

Samsung Ends Software Updates for Galaxy Z Flip 5G and Tab S7