OpenAI has released a significant update to Codex, introducing features such as background computer use, an in-app browser, image generation, memory, expanded automations, and more than 90 new plugins.
The update is currently rolling out to Codex desktop app users signed into ChatGPT on macOS. Availability of computer use for EU and UK users and personalization options for Enterprise and Edu accounts are expected soon.
Codex Computer Use on Mac and Built?In Browser
Codex now has the ability to operate applications on a Mac by seeing, clicking, and typing with its own cursor, all without needing an API integration. Multiple agents can run simultaneously without disrupting the user’s ongoing work. OpenAI says this feature is useful for testing and iterating on frontend changes, evaluating applications, and working with tools that don’t offer an API.
The Codex app now features a built-in browser that allows users to comment directly on web pages, providing precise instructions to the agent. Currently, this tool is aimed at frontend and game development on localhost. OpenAI has also announced plans to extend browser control to include more than just local web applications over time.
New Image Tools, Memory, and Automation Powers
Codex now has the ability to generate and refine images using gpt-image-1.5, integrating screenshots and code. This feature can be used for creating product concept visuals, designing frontend mockups, and developing game assets within the same workflow.
A memory preview helps Codex remember context from previous sessions, including personal preferences, corrections, and information that took significant time to gather. The goal of memory is to make repeated tasks quicker to set up and to improve output quality without needing detailed custom instructions.
Automations have also been expanded, allowing reuse of existing conversation threads. Codex can now schedule future work on its own and resume long-running tasks automatically, even across multiple days or weeks.
According to OpenAI, teams are using automations to manage open pull requests, follow up on tasks, and monitor activity across Slack, Gmail, and Notion. Additionally, Codex now proactively suggests work based on the project’s context, connected plugins, and memory. This includes identifying open comments in Google Docs, pulling relevant information from connected tools, and providing a prioritized list of actions.
Developer Workflow Improvements and 90+ Codex Plugins
The app now allows users to respond directly to GitHub review comments, run multiple terminal tabs at once, and connect to remote devboxes via SSH in the alpha version. Files can be opened directly from the sidebar with detailed previews for PDFs, spreadsheets, slides, and documents. A new summary pane provides an overview of agent plans, sources, and artifacts.
More than 90 additional plugins have been added, expanding support for skills, app integrations, and MCP servers. The new plugins include Atlassian Rovo for Jira management, CircleCI, CodeRabbit, GitLab Issues, Microsoft Suite, Neon by Databricks, Remotion, Render, and Superpowers.
Thank you for being a Ghacks reader. The post OpenAI Updates Codex With Computer Use, In-App Browser, Memory, and 90-Plus New Plugins appeared first on gHacks.
